Compare commits

..
Author SHA1 Message Date
Hermenix 5177b4c47d tests: keep the pixel buffer alive instead of taking an address of a temporary
The sandbox caught this and the local run did not: ctypes.create_string_buffer in
an expression is collected as soon as the statement ends, so the address handed to
the encoder dangles and the memory is free to be reused. Locally the freed bytes
still held the raster; under the build's allocator they did not, and the encoder
was handed garbage -- 'block (0, 0) came back as (27, 0, 30)'.

Same class of bug as the two the encoder itself documents, which is why the helper
now owns the buffer and says so.
2026-10-11 23:07:15 +00:00
Hermenix b2bc84602a build: point TURBOJPEG_PATH at lib.getLib, not the package's default output
The default output of libjpeg-turbo in this nixpkgs is the tools (-bin), which has
no lib/ at all, so the check failed with 'cannot load libjpeg-turbo from
.../-bin/lib/libturbojpeg.so'. lib.getLib selects the output that actually holds
the shared library.
2026-10-11 23:04:17 +00:00
Hermenix a417a6919d tests: find libjpeg-turbo by TURBOJPEG_PATH, not LD_LIBRARY_PATH
The unit check failed with the library present:  in nativeBuildInputs is a
nix wrapper that sets LD_LIBRARY_PATH itself, so an export from the builder script
is overwritten before the tests run. TURBOJPEG_PATH is read first by
jpeg._load_library, no wrapper touches it, and the package's console scripts now
set it too.
2026-10-11 23:01:25 +00:00
Hermenix 7190960120 capture: encode JPEG in-process too, so the live stream keeps working
PNG-only was a real defect, not a nice-to-have: `hhstream_live.server` serves
`multipart/x-mixed-replace` with `image/jpeg` parts and its own capture default is
`grim -c -t jpeg -q 80 -`, so switching this backend on would have broken video
streaming while looking faster in a benchmark.

JPEG now goes through libjpeg-turbo by ctypes, using its `TJPF_BGRX` input format
-- which is exactly the compositor's B, G, R, X memory order, so the frame goes
into the encoder with no copy and no colour conversion. Measured on this VM
against the real compositor, the deployment's own settings (Virtual-1, 1280x720,
jpeg q80, 200 frames each, two runs):

                        p50 ms   fps    our CPU   compositor CPU
    in-process           76.2   13.1     2.1 ms      51.5 ms
    grim -c -t jpeg     137.0    7.3     0.5 ms      80.2 ms
                                                          (second run: 84/144,
                                                           55.1/79.8 -> 1.71x)

So ~1.7-1.8x the frame rate and ~1.5x less of the compositor's CPU, with frames
of the same order (188 KB against grim's 222 KB). PNG stays available through the
stdlib for a deployment that would rather not carry the library.

Both segfault-class traps from the spike are handled and documented: every
argtype is declared (an undeclared pitch is a crash, not a wrong number), and
`*jpegBuf` is reset to NULL after `tjFree` (turbojpeg reads a non-NULL pointer as
a caller buffer, so the next frame would write into freed memory). The decoder is
here because a lossy codec can only be tested by round-tripping pixels -- the new
suite encodes solid blocks whose channels differ enough that a swap is
unmistakable, and includes a padded stride.

Packaging: libjpeg-turbo is reached through ctypes, so nothing links it and the
path must reach the loader -- the console scripts get a wrapper with it on
LD_LIBRARY_PATH, and the unit check exports it too. A missing library is reported
at startup rather than once per frame: a server that is up but can never produce
a frame is worse than one that refuses to start.

Verified end to end: server with `--capture wayland:Virtual-1 --format jpeg`,
client Hello `image/jpeg`, three frames read, each magic `ffd8`.
2026-10-11 22:58:28 +00:00
Hermenix a813e0b4f8 protocol: spell the system out in regenerate.py's recipe
The docstring told the next reader to use ${builtins.currentSystem}, which a shell
expands before nix ever sees it -- the expression then fails to parse with
'syntax error, unexpected ;'. It also omitted the two environment variables the
script requires (HHSTREAM_WAYLAND_XML and PYTHONPATH), so the recipe as written
could not have worked. Both fixed, with the reason recorded.
2026-10-11 22:47:12 +00:00
Hermenix a7d9357b43 capture: an in-process Wayland backend, and the measurements that justify it
The frame server could only capture by running a program once per frame
(grim). That works, but ~95% of a grim frame is waiting -- fork, connect,
handshake, round trip -- not encoding, and grim makes the compositor import
a *new* shared-memory buffer for every single frame.

This adds `--capture wayland[:OUTPUT]`: one long-lived zwlr-screencopy-v1
client per server, with one shm buffer reused for the whole run, capturing
inside the process. Measured here at 1280x720 (spike capture-speed, 200
frames; live re-check on Virtual-1, 60 frames):

    grim (subprocess)    117.3 ms/frame   8.1 fps   compositor 116.4 ms
    grim-rs (Rust crate) 124.6 ms/frame   7.5 fps   compositor 123.1 ms
    this backend          78.0 ms/frame  13.9 fps   compositor  67.5 ms

    live, PNG both sides: this 128.5 ms/frame against grim's 389.3 ms (3.0x)

Three findings, each of which shaped the code:

* The win is buffer reuse, not language. grim and grim-rs both hand the
  compositor a fresh buffer per frame, which is why their compositor CPU is
  nearly double ours. The per-frame allocate-copy-swizzle that grim-rs pays
  was measured in isolation at 2.4 ms of *client* time; the rest of its
  deficit lands inside the compositor.
* ext-image-copy-capture-v1 is no faster (78.0 vs 77.7 ms) and its
  create_frame path appears in a Hyprland segfault stack, so this keeps the
  deprecated protocol grim uses. The switch is one constant when that crash
  is fixed.
* PNG only, deliberately: PNG is the one encoding this package produces
  without a dependency, so --format jpeg is refused rather than publishing a
  Hello whose MIME type the frames do not have. A PNG frame is ~2.1 MB
  against ~120 KB for JPEG q80, which is why the command backend stays the
  default and why a libjpeg-turbo encoder (1.7 ms/frame, measured) is the
  obvious follow-up.

The timeout is the load-bearing part of the client: pywayland only reads
events from read()/dispatch(block=True), neither of which takes a timeout,
and the server captures in a worker thread where SIGALRM is unavailable. It
therefore polls the display fd itself, so a compositor that stops answering
produces a CaptureError instead of a capture thread that never returns.

The protocol bindings are vendored under src/hhstream/_protocol/ exactly the
way hhstream_pointer already vendors its two, with regenerate.py recording
where each XML came from. No new dependency: pywayland was already required
by the input driver.

Contract: the new server flag and the two new module options are documented
in docs/CONTRACT.md, and docs/ARCHITECTURE.md's prediction that a native
screencopy client would be "a swap rather than a rewrite" is now the record
of what the swap cost.

Unit suite: 977 passed, 2 skipped, 0 failed (nix build -L .#checks.unit).
The live numbers were taken on this VM against the real compositor.
2026-10-11 20:34:33 +00:00
346 changed files with 1533 additions and 10936 deletions

No files matched your search

-136
View File
@@ -1,136 +0,0 @@
# The measurement contract — frozen
Two people are building capture producers in two worktrees. If each invents its
own timing, the comparison is worthless. So this file is the interface, and it is
**frozen**: implement to it exactly, do not improve it.
## The producer command
Every route is a command that captures frames and reports its own timings as one
JSON object. Same argv, same output, whatever language it is written in.
```
<producer> --output <wl_output_name> --frames <N> --warmup <W> \
--mode {raw|jpeg} --json
<producer> --all --frames <N> --warmup <W> --mode {raw|jpeg} --json
```
| Flag | Meaning |
|---|---|
| `--output NAME` | capture exactly this `wl_output` (e.g. `hh-<session-id>`, or the main output) |
| `--all` | capture **every** output and stitch them, left to right, into one image — this is the 2560×720 geometry `hhstream` publishes today. Mutually exclusive with `--output`. |
| `--frames N` | **measured** frames (after warm-up) |
| `--warmup W` | discarded frames, taken first, on the same connection |
| `--mode raw` | the frame is complete in a userspace buffer; **no encode, no format conversion** |
| `--mode jpeg` | the frame is complete as JPEG bytes |
| `--json` | print the result object (below) as the **last line of stdout** |
Anything else the producer wants to say goes to **stderr**. Stdout carries
exactly one JSON object, so a runner can parse it without guessing.
## What is timed
**One frame = from just before the capture is requested, to the moment that
frame's bytes are complete in the producer's own userspace buffer.**
- `--mode raw`: stop when the compositor's frame is ready and the buffer is
mapped and readable. Do **not** convert ARGB→RGB, do not apply transforms, do
not touch the pixels. Reads that a consumer would do (stride, checksum) are
outside the measurement.
- `--mode jpeg`: stop when the encoded bytes are complete in the buffer, however
you encode them (library, subprocess, hardware). Encoding is inside the
measurement, because encoding is what a live stream pays.
- For the **subprocess baseline** (`grim`), the timed span is the whole
`fork`+`exec`+run+read: spawn, write to a pipe, read to EOF. That is the honest
cost of the route we are trying to beat, and the fork cost is the point.
- Exclude warm-up frames from every statistic. Use a monotonic clock.
- The first frame after a connection is established is a warm-up frame by
definition — allocate buffers and settle the protocol before measuring.
## The result object
```json
{
"route": "grim-subprocess", // short id of the implementation
"language": "python", // python | rust
"protocol": "zwlr-screencopy-v1", // what actually served the frames
"output": "hh-chromium-yrfadn", // or "all"
"geometry": [1280, 720],
"mode": "jpeg",
"warmup": 10,
"frames": 300,
"frame_ms": { "min": 0.0, "p50": 0.0, "mean": 0.0, "p95": 0.0, "max": 0.0 },
"fps_sustained": 0.0, // frames / (sum of frame_ms / 1000)
"wall_ms": 0.0, // total wall time of the measured run
"cpu_ms_per_frame": null, // process CPU time / frames, or null if unknown
"bytes_p50": null, // frame size in bytes (jpeg mode), else null
"persistent": true, // one long-lived connection, or a process per frame
"notes": "" // anything a reader must know to read the numbers
}
```
### Resource fields (added 2026-10-11 — the operator's requirement)
Speed alone does not decide this. A route that hits 15 fps by burning a core is
not a route we can ship on a machine that also renders. Five more fields are
**required**, measured over the same measured window as `frame_ms`:
```json
"rss_peak_mb": 0.0, // peak RSS of the producer (children included for subprocess routes)
"rss_peak_kb_raw": 0, // the unrounded value, so nobody re-derives a rounded one
"system_cpu_ms_per_frame": 0.0, // every process on the machine, from /proc/stat
"compositor_cpu_ms_per_frame": 0.0, // the compositor's own utime+stime, from /proc/<hyprland>/stat
"machine_cpu_pct": 0.0, // system_cpu over the window, as % of one core (can exceed 100)
"frames_bytes_allocated_p50": null, // bytes of buffer the route holds per frame (raw mode: the whole frame)
```
- `system_cpu_ms_per_frame` is the honest "what this costs the machine" number,
and it is the one that catches a route which is fast only because the
compositor did extra work. Report it even when it is unflattering.
- `compositor_cpu_ms_per_frame` separates "the capture is expensive" from "the
capture made the compositor expensive". Read `/proc/<pid>/stat` fields 14+15
(utime+stime, in clock ticks — divide by `os.sysconf("SC_CLK_TCK")`). If the
file is unreadable, say so in `notes` rather than reporting zero.
- `rss_peak_mb`: `resource.getrusage(RUSAGE_SELF).ru_maxrss` for a persistent
route; `RUSAGE_CHILDREN` for a subprocess-per-frame route (it is the peak of
the children, which is what actually matters there). ru_maxrss is in **KiB**
on Linux, despite the name.
- These are measured *around the measured loop only* — warm-up excluded — so
they line up with `frame_ms`.
### The conditions a run is reported under
- `protocol` must name what really ran, not what you hoped: if
`ext-image-copy-capture-v1` is unavailable and you fell back to
`zwlr-screencopy-v1`, say so. If you had to fall back at runtime, say that too.
- `persistent` separates "one connection, N frames" from "a process per frame" —
the two are not comparable and the table must not mix them silently.
- `cpu_ms_per_frame` is process CPU (`/proc/self/stat` utime+stime, or
`getrusage`) divided by frames. A route that is fast by pinning a core is not
fast; report it.
- Numbers with no measurement behind them stay `null`. Never estimate a field.
## The conditions a run is reported under
The runner runs every route under identical conditions and records them:
1. **Geometry**: `--output <one container>` (1280×720) and `--all` (2560×720).
2. **Screen**: a *still* pass and a *moving* pass. Motion is made by a session
printing in a loop or a page being scrolled; the runner starts it, waits for
it to be visibly moving, and destroys it afterwards. A still screen can make a
damage-driven capture wait — if that happens, it is a **finding** and goes in
the verdict, not in a footnote.
3. **Mode**: `raw` and `jpeg`.
4. **Repetition**: every cell twice; report both runs, not just the better one.
## The rule that keeps this honest
**The baseline is `grim` as it runs today.** Whatever beats it, beats that, and
the verdict has to say by how much — or that nothing did. A route that is only
faster in a mode the live stream does not use (raw pixels, when the wire carries
JPEG) must be labelled as such rather than presented as a win.
And the environment caveat that goes in every verdict: **this VM has no GPU.**
Hyprland renders through software (kms_swrast / llvmpipe). Zero-copy dma-buf
paths cannot help here, so a route must not be tuned to take one; say what a
GPU machine would change, but do not claim it as measured.
-132
View File
@@ -1,132 +0,0 @@
# SPIKE BRIEF — two capture routes, measured head to head
An experiment, in a fresh worktree: `/var/lib/hermes/worktrees/hhs-capture-speed`,
branch `spike/capture-speed`, forked from `main` at `f3ac9ad`. Read
`README.dev.md` and `docs/CONTRACT.md` first; the repo's conventions apply
(docstrings that say *why*, `nix develop`, no CLI shims where a socket will do).
This is a **spike**, not production: `spikes/` is throwaway code, but the numbers
must be real. See the `spike` convention: decompose → research → build → verdict.
## The question the operator asked
> Which approach lets us render images faster — the existing `grim` subprocess, a
> Rust library (`grim-rs`), or a raw Python Wayland client? This is for **live
> streaming the workspaces to the web, and for agent eyes.**
"Render images faster" means, concretely: **time from "wake up and ask" to
"pixels/JFIF bytes in a buffer I can publish"**, sustained, and the frame rate
that follows. Encode cost is part of it, because what gets published is JPEG.
## The three routes
Measure all three with **one harness**, in this order (baseline first, so the
comparison is anchored):
- **000 — baseline: `grim` as a subprocess.** The shipped path:
`grim -c -t jpeg -q 80 -o <output> -` per frame. One process per frame.
A prior measurement in this repo says ~46 ms/frame at 1280x720 q80 **— do not
trust it, re-measure it here.**
- **001a — `grim-rs`.** The Rust crate (crates.io `grim-rs`, ~0.2): native
capture over `ext-image-copy-capture-v1` with a `zwlr_screencopy_manager_v1`
fallback, PNG/JPEG encoders included. Measure **two variants separately** and
label which number came from which:
- **(a1)** a long-lived Rust producer process — one connection, N frames
captured and encoded in a loop (the "another grim, but persistent" model);
- **(a2)** in-process from Python via PyO3/maturin, **if it fits in the
spike**. If it does not, say so plainly and leave a1 as the number. Do not
fake a2.
- **001b — raw Python.** `pywayland` + vendored protocol XML for
`ext-image-copy-capture-v1` (Hyprland has supported it since 0.52.1; this
machine runs 0.56.2 and ships `src/protocols/ImageCopyCapture.cpp`), with a
`zwlr_screencopy` fallback if the newer protocol turns out to be unavailable
in practice. **Follow the precedent already in this repo**: the input driver
vendors its protocol bindings under `src/hhstream_pointer/_protocol/` and
documents how to regenerate them (`regenerate.py`) — do the same in the spike,
and say where you got the XML.
Research before building (that is the point of the research step): read the
`grim-rs` API and the `ext-image-copy-capture-v1` spec, and check what
`pywayland` needs to bind a new protocol. Pick approaches and say why.
## Environment facts — get these right or the numbers are meaningless
- The compositor is **headless Hyprland 0.56.2 running as a system service**, not
in your shell. The published Wayland socket is
**`WAYLAND_DISPLAY=/run/hypr-stream/wayland-1`** (an absolute path).
**Do NOT set `XDG_RUNTIME_DIR=/run/hypr-stream`** — that directory is owned by
the `hhstream` user and libwayland refuses a runtime dir the caller does not
own; the failure reads as "failed to create display".
- The frame socket is `/run/hypr-stream/stream.sock` (`hhstream status` reports
all of this). `hhstream shot --output <name>` is the quickest way to prove a
capture path works before you time it.
- **There is no GPU.** The compositor renders through software (kms_swrast /
llvmpipe). So: **do not use the zero-copy dma-buf / GPU paths to make the Rust
route look good** — they cannot help on this box, and a benchmark that silently
takes a GPU path and falls back would not be honest. Measure the CPU path, and
say in the verdict what a GPU box would change.
- Geometries to measure (both, always): **1280×720** (one session container) and
**2560×720** (the whole desktop as `hhstream` publishes it today — this is the
geometry on the wire now).
- Measure **a moving screen and a still one.** A still screen with damage-driven
capture can block or go quiet, which is exactly the kind of thing that decides
a live-stream design — the current `hhstream` path re-encodes the same frame on
a timer instead. If a route behaves differently on a still screen, that is a
finding, not a nuisance. To make motion: start a session
(`hhstream session start chromium -- --ozone-platform=wayland --new-window
https://example.com` — the allowlist is in the NixOS module) or use
`examples/agent-eyes-demo.py`, which already drives a page. A `foot` session
with something printing in a loop is cheaper and moves plenty.
- A live session for testing may already exist (`hhstream session list`); you may
start your own. **Do not kill another agent's session without saying so.**
## The harness (build this once, use it for all routes)
- Warm-up frames excluded; then **≥300 measured frames per cell**; report
**p50 / p95 / max per-frame ms**, sustained fps, and **CPU time per frame**
(`time.process_time` / `/proc` for the subprocess cases) — a route that is fast
by pinning a core is not fast.
- Separate the two stages so the verdict can say where the time goes:
**(i) frame ready in a buffer**, **(ii) + JPEG encode**. Report both. If a
route cannot separate them, say so.
- Same output, same geometry, same content for every route; run each cell at
least twice and report the spread, not one lucky run.
- Print a machine-readable line per cell (JSON) **and** a human table, and write
the raw runs to a file so the numbers in the README can be re-derived.
## Deliverables
1. `spikes/capture-speed/README.md` — the question, the method, the head-to-head
table, and a `## Verdict` (VALIDATED / PARTIAL / INVALIDATED) with what worked,
what didn't, surprises, and a recommendation for the real build. The existing
`grim` route and the two new ones all get a row.
2. `spikes/capture-speed/001a-grim-rs/README.md` and `001b-pywayland/README.md`
with their own verdicts and the raw evidence.
3. **A chart PNG** (frame time and sustained fps per route, both geometries).
`ad-hoc-charts` conventions if you want a recipe; matplotlib via
`nix shell nixpkgs#python3Packages.matplotlib` is fine.
4. **Something the operator can watch** — this matters more than the table. The
cheapest honest version: run each *winning-capable* backend behind an MJPEG
endpoint the operator can open in a browser, both at once or one at a time,
with the achieved fps printed on the page or in its sidecar `status.json`.
Reuse `src/hhstream_live/` (its publisher takes a capture command/loop) rather
than writing a second web app; if reuse turns out to be the harder path, say
so and keep the demo small. Bind it sensibly: the guest is in NAT mode, so a
browser on the host may or may not reach a port — say what you bound and how
you verified it with `curl`, and do not touch the system firewall.
5. A short "what a production version would take" section: which file in
`src/hhstream/` the winner would replace, what dependencies it adds, and what
the repo's stdlib-only policy would cost.
## Rules
- **Do not touch the running system**: no `nixos-rebuild`, no edits under
`/etc/nixos`, no changes to the running `hhstream` units. You may run the CLI
and start/stop your own sessions.
- Keep Rust out of the repo's main flake: a spike-local `shell.nix`/flake under
`spikes/capture-speed/` is the place for `cargo`, `rustc`, `maturin`.
`cargo`/`rustc` 1.98 are in nixpkgs and crates.io is reachable from here.
- Commit on `spike/capture-speed` as you go. Do not merge, do not touch `main`.
- Every number in a README must be reproducible by a command you paste next to
it. If something did not work, it goes in the README as a finding, not a
silent omission.
+16
View File
@@ -13,6 +13,7 @@
# the suite runs in a build.
{
lib,
libjpeg_turbo,
runCommand,
python3,
hhstream,
@@ -26,6 +27,8 @@ let
# whether a bad string types a prefix.
pytestEnv = python3.withPackages (ps: [ ps.pytest ps.pywayland ps.xkbcommon ]);
# Only the suite itself, so an editor's droppings or a local .pytest_cache
# cannot change the derivation's hash.
testSrc = lib.cleanSourceWith {
@@ -62,6 +65,19 @@ runCommand "hhstream-unit"
# than letting pytest do it makes a packaging problem distinguishable from a
# test failure.
export PYTHONPATH=${hhstream}/${python3.sitePackages}
# `tests/unit/test_jpeg.py` round-trips a real frame through libjpeg-turbo: a
# lossy codec can only be checked by comparing pixels, so the suite needs the
# library -- and it needs the *loader* to find it, because the encoder reaches
# it through `ctypes` and nothing links it. Same reason nix/package.nix wraps
# the console scripts with this path.
#
# `TURBOJPEG_PATH` decides, not `LD_LIBRARY_PATH`: the `pytest` in
# `nativeBuildInputs` is a nix wrapper that sets `LD_LIBRARY_PATH` itself, so
# exporting it here had no effect and the suite failed with the library plainly
# present on the machine. `jpeg._load_library` reads this variable first.
export LD_LIBRARY_PATH=${libjpeg_turbo}/lib
export TURBOJPEG_PATH=${lib.getLib libjpeg_turbo}/lib/libturbojpeg.so
export PYTHONDONTWRITEBYTECODE=1
export HOME=$PWD
python3 -c 'import hhstream; print("testing hhstream", hhstream.__version__, "from", hhstream.__file__)'
+26 -2
View File
@@ -87,8 +87,32 @@ Capturing is a separate program (by default `grim`, a wlr-screencopy client) run
inside the compositor's session and pointed at stdout. That choice is what keeps
`hhstream` free of Wayland and GPU code — the awkward part lives in a tool that
already exists and is already tested — and it is why the capture command is a
module option: tests inject a fake, and a future backend (a native screencopy
client, a hardware encoder) is a command-line swap rather than a rewrite.
module option: tests inject a fake, and a different backend is a swap rather than
a rewrite.
That swap has now happened once. `captureBackend = "wayland"` captures in-process
(`src/hhstream/wayland_capture.py`): one long-lived `zwlr-screencopy-v1` client
with **one shared-memory buffer reused for every frame**, instead of a `grim`
process per frame. Measured at 1280×720 over 200 frames, that is ~1.5× the frame
rate (78 ms against 117 ms per frame) and — the number that matters on this
machine — roughly **half the compositor's CPU per frame** (67 ms against 116 ms),
because `grim` makes the compositor import a new buffer every time. It is
deliberately the *deprecated* protocol: `ext-image-copy-capture-v1` measured
identically and appears in a Hyprland segfault stack (`create_frame` →
`CScreenshareFrame::transform`). The price of the in-process route is that the
package now carries Wayland protocol bindings of its own (`src/hhstream/_protocol/`,
vendored the same way `hhstream_pointer` already vendors its two) and that it can
publish **JPEG** (through libjpeg-turbo) as well as PNG — and JPEG is the one
that matters, because the live viewer serves `image/jpeg` parts and a PNG-only
backend would have broken video streaming while looking faster on a benchmark.
The library is reached through `ctypes` and has to be on `LD_LIBRARY_PATH`, which
the package's wrapper arranges. Both costs are why `command` is still the default:
switching is the operator's call, not something a merge should decide.
Comments in the spike that produced these numbers live in the branch's
`capture-speed` worktree; the honest caveat that came with them is that this box
is saturated by *rendering* whatever is on screen, so the win is in how much room
the compositor keeps, not in headroom that did not exist.
## The hardware requirement, stated plainly
+14 -2
View File
@@ -93,7 +93,9 @@ Namespace: `services.hyprland-headless-stream`
| `extraConfig` | str | `""` | extra lines appended to `hyprland.conf` |
| `hyprlandPackage` | package | `pkgs.hyprland` | which compositor |
| `package` | package | the flake's own build | the frame server/client |
| `captureCommand` | list of str | `[ "grim" … ]` | override the capture backend — used by tests |
| `captureCommand` | list of str | `[ "grim" … ]` | override the capture *command* — used by tests |
| `captureBackend` | `"command"` \| `"wayland"` | `"command"` | capture by command or in-process (`zwlr-screencopy-v1`); both `format` values work |
| `captureOutput` | str or null | `null` | output to capture by name when `captureBackend = "wayland"`; null = the first one |
Services (names are part of the contract, tests refer to them):
@@ -109,7 +111,8 @@ status as the wire format. Exactly this shape, and no others:
```
hhstream-server --socket <socketPath> --fps <frameRate> --format <png|jpeg> \
[--capture-command <token> ...]
[--capture-quality <1..100>] \
[--capture-command <token> ...] [--capture wayland[:NAME]]
```
- `--format` takes the **short name** (`png`, `jpeg`), not a MIME type; the
@@ -119,6 +122,15 @@ hhstream-server --socket <socketPath> --fps <frameRate> --format <png|jpeg> \
--capture-command -`. A single shell string cannot carry a leading `-`
unambiguously. The `--capture-command=<token>` form is accepted as well, and
the two may be mixed. When absent, the server derives `grim -t <format> -`.
- `--capture` selects a **named backend** instead of a program: `wayland`
captures the compositor's first output in-process, `wayland:NAME` the output
Hyprland calls `NAME`. It is exclusive with `--capture-command` (the backend
replaces the command; passing both is a caller error). It honours both
`--format` values: `png` through the stdlib, `jpeg` through libjpeg-turbo. If
libjpeg-turbo is missing and `--format jpeg` was asked for, the server **refuses
to start** rather than serving a socket it can never produce a frame for.
- `--capture-quality` is the JPEG quality for the in-process backend (1..100,
default 80, matching the default command's `-q 80`); `--format png` ignores it.
- There is deliberately **no** `--width` / `--height`: the Hello reports the
captured frame's real size (see above).
- The socket's **mode** (0660) is the server's business; its **ownership** is
+42 -1
View File
@@ -244,7 +244,13 @@ let
# shape that survives that without inventing a quoting convention.
# `hhstream.server.build_parser()` is still a stub, so this is the one
# place the module guesses at the server's CLI rather than reading it.
++ lib.concatMap (word: ["--capture-command" word]) cfg.captureCommand;
++ (
if cfg.captureBackend == "wayland"
then ["--capture" (
if cfg.captureOutput == null then "wayland" else "wayland:${cfg.captureOutput}"
)]
else lib.concatMap (word: ["--capture-command" word]) cfg.captureCommand
);
# Shared by both units: a private session, no parent display to nest in.
sessionEnvironment = {
@@ -398,6 +404,41 @@ in {
description = "Command that writes one frame to stdout; overridable for tests.";
};
captureBackend = lib.mkOption {
type = lib.types.enum ["command" "wayland"];
default = "command";
description = ''
How frames are obtained: `command` runs {option}`captureCommand` once per
frame; `wayland` captures in-process, over the compositor's own
`zwlr-screencopy-v1`, with no subprocess at all.
Measured on a 1280x720 output (spike `capture-speed`, 200 frames), the
in-process backend is ~1.5x the frame rate of `grim` and -- the figure
that matters on a small machine -- roughly half the *compositor's* CPU
per frame, because it reuses one shared-memory buffer instead of making
`grim` build a new one per frame.
Both of the {option}`format` values work, and in practice this means
`jpeg`: the live viewer serves `image/jpeg` parts, so a PNG-only backend
would break video streaming. JPEG is encoded in-process through
libjpeg-turbo at the same quality the default `grim` command uses
(~1.7 ms a frame, ~120 KB), and PNG remains available through the stdlib
for a deployment that would rather not carry that library (~2.1 MB a
frame). The default is still `command` only because switching backends
is the operator's call, not because anything is missing.
'';
};
captureOutput = lib.mkOption {
type = lib.types.nullOr lib.types.str;
default = null;
description = ''
Which output to capture by name (Hyprland's monitor name, e.g.
`Virtual-1`) when {option}`captureBackend` is `wayland`. `null` means the
compositor's first output, which is what `grim` with no `-o` uses.
'';
};
sessions = {
enable = lib.mkOption {
type = lib.types.bool;
+28 -1
View File
@@ -1,5 +1,6 @@
{
lib,
libjpeg_turbo,
python3,
...
}:
@@ -22,7 +23,26 @@ python3.pkgs.buildPythonApplication {
# hhstream_pointer._protocol.regenerate for how to refresh them -- including
# the note that virtual-keyboard-unstable-v1.xml is not in wayland-protocols
# or wlr-protocols in this nixpkgs and comes from pkgs.hyprland.src.
dependencies = [ python3.pkgs.pywayland python3.pkgs.xkbcommon ];
dependencies = [ python3.pkgs.pywayland python3.pkgs.xkbcommon libjpeg_turbo ];
# The in-process jpeg encoder reaches libjpeg-turbo through `ctypes`, so the
# library has to be *findable at run time*, which a propagated build input
# alone does not achieve: nothing links it, so nothing puts its directory on
# LD_LIBRARY_PATH. The wrapper is what makes `ctypes.util.find_library` succeed
# inside the unit -- and it is the same library grim already links, so the
# frames stay byte-compatible with the existing stream.
#
# Both variables, deliberately. `LD_LIBRARY_PATH` is the conventional one, but
# a nix Python wrapper sets that variable *itself* -- which is how the unit
# check failed with the library plainly present -- so `TURBOJPEG_PATH` is the
# one that actually decides, and `jpeg._load_library` checks it first.
makeWrapperArgs = [
"--prefix LD_LIBRARY_PATH : ${lib.makeLibraryPath [ libjpeg_turbo ]}"
# `lib.getLib`, not the bare package: the default output of libjpeg-turbo in
# this nixpkgs is the *tools* (`-bin`), which has no `lib/` at all -- pointing
# at it produced "cannot load libjpeg-turbo from .../-bin/lib/libturbojpeg.so".
"--set TURBOJPEG_PATH ${lib.getLib libjpeg_turbo}/lib/libturbojpeg.so"
];
# Unit tests are wired up in the flake's `checks` so they can run without
# installing the package; see docs/CONTRACT.md.
@@ -38,6 +58,13 @@ python3.pkgs.buildPythonApplication {
# The web app: importing the server pulls in the router, the publisher and
# the pages.
"hhstream_live.server"
# The capture backend and its vendored protocol bindings: importing the
# backend runs the scanner-generated `_gen_c()` for wl_shm, wl_output and
# zwlr-screencopy-manager-v1, which is the same cheap proof the pointer
# driver gets above.
"hhstream.wayland_capture"
"hhstream._protocol.wlr_screencopy_unstable_v1"
"hhstream.jpeg"
];
# `pythonImportsCheck` only *imports*, and `hhstream_live.phrase` reads its
-4
View File
@@ -1,4 +0,0 @@
# The Rust build tree: reproducible from Cargo.lock, and huge.
001a-grim-rs/target/
__pycache__/
generated/**/__pycache__/
@@ -1 +0,0 @@
/target
File diff suppressed because it is too large. Load diff
@@ -1,16 +0,0 @@
[package]
name = "grimrs-capture"
version = "0.1.0"
edition = "2021"
publish = false
description = "Long-lived grim-rs capture producer for the frozen HARNESS.md contract"
[[bin]]
name = "grimrs-capture"
path = "src/main.rs"
[dependencies]
# Exact: the raw-mode stop point depends on what 0.2.0 does inside capture_*.
grim-rs = "=0.2.0"
libc = "0.2"
serde_json = { version = "1", features = ["preserve_order"] }
@@ -1,53 +0,0 @@
#!/bin/sh
# Eight measured cells, still screen, two runs each.
# Stdout of each cell is one JSON line. Progress goes to runs/progress.log.
# Does not set XDG_RUNTIME_DIR. The caller must export WAYLAND_DISPLAY to the
# absolute socket path.
set -u
cd "$(dirname "$0")"
if [ "${WAYLAND_DISPLAY:-}" != "/run/hypr-stream/wayland-1" ]; then
echo "WAYLAND_DISPLAY must be /run/hypr-stream/wayland-1" >&2
exit 2
fi
if [ "${XDG_RUNTIME_DIR:-}" = "/run/hypr-stream" ]; then
echo "refusing XDG_RUNTIME_DIR=/run/hypr-stream" >&2
exit 2
fi
OUTPUT=${1:?usage: run-cells.sh OUTPUT_NAME}
BIN=./target/release/grimrs-capture
if [ ! -x "$BIN" ]; then
echo "missing $BIN; build with nix-shell --run 'cargo build --release'" >&2
exit 2
fi
mkdir -p runs
LOG=runs/progress.log
: > "$LOG"
# The frame server is a system unit. Do not stop it. Record that it was
# capturing during the run so the numbers are not read as an idle compositor.
ps -eo pid,user,cmd > runs/ps-at-start.txt
echo "BEGIN runner $(date -Is) WAYLAND_DISPLAY=$WAYLAND_DISPLAY XDG_RUNTIME_DIR=${XDG_RUNTIME_DIR-}" >> "$LOG"
fail=0
run() {
name=$1
shift
echo "BEGIN $name $(date -Is)" >> "$LOG"
start=$(date +%s)
"$BIN" "$@" > "runs/$name.stdout" 2> "runs/$name.stderr"
rc=$?
end=$(date +%s)
echo "END $name rc=$rc elapsed=$((end - start)) $(date -Is)" >> "$LOG"
if [ "$rc" -ne 0 ]; then
echo "CELL_FAIL $name rc=$rc" >> "$LOG"
fail=1
fi
}
run output-raw-1 --output "$OUTPUT" --frames 300 --warmup 10 --mode raw --json
run output-raw-2 --output "$OUTPUT" --frames 300 --warmup 10 --mode raw --json
run output-jpeg-1 --output "$OUTPUT" --frames 300 --warmup 10 --mode jpeg --json
run output-jpeg-2 --output "$OUTPUT" --frames 300 --warmup 10 --mode jpeg --json
run all-raw-1 --all --frames 300 --warmup 10 --mode raw --json
run all-raw-2 --all --frames 300 --warmup 10 --mode raw --json
run all-jpeg-1 --all --frames 300 --warmup 10 --mode jpeg --json
run all-jpeg-2 --all --frames 300 --warmup 10 --mode jpeg --json
echo "ALL_DONE fail=$fail $(date -Is)" >> "$LOG"
exit "$fail"
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol zwlr-screencopy-v1 mode jpeg target all warmup 10 frames 300
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 2/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 3/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 4/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 5/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 6/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 7/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 8/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 9/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 10/10 2560x720 bytes 284962 nonzero 281335
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"zwlr-screencopy-v1","output":"all","geometry":[2560,720],"mode":"jpeg","warmup":10,"frames":300,"frame_ms":{"min":114.651,"p50":196.918,"mean":207.359,"p95":304.217,"max":471.059},"fps_sustained":4.823,"wall_ms":62207.988,"cpu_ms_per_frame":13.017,"bytes_p50":284962.0,"persistent":true,"notes":"grim-rs 0.2.0 via Grim::new_wlr(); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 281335."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol zwlr-screencopy-v1 mode jpeg target all warmup 10 frames 300
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 2/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 3/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 4/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 5/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 6/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 7/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 8/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 9/10 2560x720 bytes 284962 nonzero 281335
grimrs-capture: warmup 10/10 2560x720 bytes 284962 nonzero 281335
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"zwlr-screencopy-v1","output":"all","geometry":[2560,720],"mode":"jpeg","warmup":10,"frames":300,"frame_ms":{"min":115.964,"p50":185.942,"mean":200.869,"p95":280.883,"max":344.797},"fps_sustained":4.978,"wall_ms":60261.001,"cpu_ms_per_frame":13.072,"bytes_p50":284962.0,"persistent":true,"notes":"grim-rs 0.2.0 via Grim::new_wlr(); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 281335."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol zwlr-screencopy-v1 mode raw target all warmup 10 frames 300
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 2/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 3/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 4/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 5/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 6/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 7/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 8/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 9/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 10/10 2560x720 bytes 7372800 nonzero 7353441
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"zwlr-screencopy-v1","output":"all","geometry":[2560,720],"mode":"raw","warmup":10,"frames":300,"frame_ms":{"min":105.174,"p50":179.929,"mean":191.042,"p95":269.35,"max":486.399},"fps_sustained":5.234,"wall_ms":57312.833,"cpu_ms_per_frame":2.41,"bytes_p50":null,"persistent":true,"notes":"grim-rs 0.2.0 via Grim::new_wlr(); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 7353441."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol zwlr-screencopy-v1 mode raw target all warmup 10 frames 300
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 2/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 3/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 4/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 5/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 6/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 7/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 8/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 9/10 2560x720 bytes 7372800 nonzero 7353441
grimrs-capture: warmup 10/10 2560x720 bytes 7372800 nonzero 7353441
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"zwlr-screencopy-v1","output":"all","geometry":[2560,720],"mode":"raw","warmup":10,"frames":300,"frame_ms":{"min":100.09,"p50":175.426,"mean":189.164,"p95":267.965,"max":595.464},"fps_sustained":5.286,"wall_ms":56749.543,"cpu_ms_per_frame":2.415,"bytes_p50":null,"persistent":true,"notes":"grim-rs 0.2.0 via Grim::new_wlr(); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 7353441."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol zwlr-screencopy-v1 mode jpeg target hh-foot-utjcf2 warmup 10 frames 300
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 2/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 3/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 4/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 5/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 6/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 7/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 8/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 9/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 10/10 1280x720 bytes 92861 nonzero 90786
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"zwlr-screencopy-v1","output":"hh-foot-utjcf2","geometry":[1280,720],"mode":"jpeg","warmup":10,"frames":300,"frame_ms":{"min":43.254,"p50":119.493,"mean":126.816,"p95":212.019,"max":329.716},"fps_sustained":7.885,"wall_ms":38044.93,"cpu_ms_per_frame":6.721,"bytes_p50":92861.0,"persistent":true,"notes":"grim-rs 0.2.0 via Grim::new_wlr(); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 90786."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol zwlr-screencopy-v1 mode jpeg target hh-foot-utjcf2 warmup 10 frames 300
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 2/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 3/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 4/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 5/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 6/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 7/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 8/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 9/10 1280x720 bytes 92861 nonzero 90786
grimrs-capture: warmup 10/10 1280x720 bytes 92861 nonzero 90786
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"zwlr-screencopy-v1","output":"hh-foot-utjcf2","geometry":[1280,720],"mode":"jpeg","warmup":10,"frames":300,"frame_ms":{"min":46.294,"p50":112.667,"mean":112.794,"p95":171.372,"max":257.709},"fps_sustained":8.866,"wall_ms":33838.267,"cpu_ms_per_frame":6.66,"bytes_p50":92861.0,"persistent":true,"notes":"grim-rs 0.2.0 via Grim::new_wlr(); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 90786."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol zwlr-screencopy-v1 mode raw target hh-foot-utjcf2 warmup 10 frames 300
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 2/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 3/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 4/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 5/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 6/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 7/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 8/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 9/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 10/10 1280x720 bytes 3686400 nonzero 3679565
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"zwlr-screencopy-v1","output":"hh-foot-utjcf2","geometry":[1280,720],"mode":"raw","warmup":10,"frames":300,"frame_ms":{"min":51.409,"p50":110.264,"mean":109.115,"p95":177.876,"max":215.254},"fps_sustained":9.165,"wall_ms":32734.584,"cpu_ms_per_frame":1.445,"bytes_p50":null,"persistent":true,"notes":"grim-rs 0.2.0 via Grim::new_wlr(); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 3679565."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol zwlr-screencopy-v1 mode raw target hh-foot-utjcf2 warmup 10 frames 300
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 2/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 3/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 4/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 5/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 6/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 7/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 8/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 9/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 10/10 1280x720 bytes 3686400 nonzero 3679565
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"zwlr-screencopy-v1","output":"hh-foot-utjcf2","geometry":[1280,720],"mode":"raw","warmup":10,"frames":300,"frame_ms":{"min":46.687,"p50":106.48,"mean":103.438,"p95":157.55,"max":262.651},"fps_sustained":9.668,"wall_ms":31031.55,"cpu_ms_per_frame":1.386,"bytes_p50":null,"persistent":true,"notes":"grim-rs 0.2.0 via Grim::new_wlr(); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-foot-utjcf2@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 3679565."}
@@ -1,134 +0,0 @@
PID USER CMD
1 root /run/current-system/systemd/lib/systemd/systemd --switched-root --system --deserialize=47
2 root [kthreadd]
3 root [pool_workqueue_release]
4 root [kworker/R-rcu_gp]
5 root [kworker/R-sync_wq]
6 root [kworker/R-kvfree_rcu_reclaim]
7 root [kworker/R-slub_flushwq]
8 root [kworker/R-netns]
10 root [kworker/0:0H-events_highpri]
13 root [kworker/R-mm_percpu_wq]
15 root [ksoftirqd/0]
16 root [rcu_preempt]
17 root [rcu_exp_par_gp_kthread_worker/0]
18 root [rcu_exp_gp_kthread_worker]
19 root [migration/0]
20 root [idle_inject/0]
21 root [cpuhp/0]
22 root [cpuhp/1]
23 root [idle_inject/1]
24 root [migration/1]
25 root [ksoftirqd/1]
27 root [kworker/1:0H-events_highpri]
28 root [kdevtmpfs]
29 root [kworker/R-inet_frag_wq]
30 root [rcu_tasks_kthread]
31 root [rcu_tasks_rude_kthread]
32 root [rcu_tasks_trace_kthread]
33 root [kauditd]
34 root [khungtaskd]
35 root [oom_reaper]
37 root [kworker/R-writeback]
39 root [kcompactd0]
40 root [ksmd]
41 root [khugepaged]
42 root [kworker/R-kblockd]
43 root [kworker/R-blkcg_punt_bio]
44 root [kworker/R-kintegrityd]
45 root [irq/9-acpi]
46 root [kworker/R-tpm_dev_wq]
47 root [kworker/R-devfreq_wq]
48 root [kworker/R-quota_events_unbound]
49 root [kworker/1:1H-kblockd]
50 root [kswapd0]
51 root [kdamond.0]
52 root [kworker/R-kthrotld]
53 root [kworker/R-mld]
54 root [kworker/R-ipv6_addrconf]
55 root [kworker/R-kstrp]
56 root [kworker/u9:0]
57 root [kworker/0:1H-kblockd]
59 root [hwrng]
126 root [kworker/R-ata_sff]
128 root [scsi_eh_0]
129 root [kworker/R-scsi_tmf_0]
130 root [scsi_eh_1]
131 root [kworker/R-scsi_tmf_1]
132 root [scsi_eh_2]
133 root [kworker/R-scsi_tmf_2]
134 root [scsi_eh_3]
135 root [kworker/R-scsi_tmf_3]
136 root [scsi_eh_4]
137 root [kworker/R-scsi_tmf_4]
138 root [scsi_eh_5]
139 root [kworker/R-scsi_tmf_5]
167 root [jbd2/vda2-8]
168 root [kworker/R-ext4-rsv-conversion]
176 root [jbd2/vdb-8]
177 root [kworker/R-ext4-rsv-conversion]
305 root /nix/store/baxgs06659w4i6ggsfr56srjrb52v1h8-systemd-261.2/lib/systemd/systemd-journald
308 systemd+ /nix/store/baxgs06659w4i6ggsfr56srjrb52v1h8-systemd-261.2/lib/systemd/systemd-oomd
330 systemd+ /nix/store/baxgs06659w4i6ggsfr56srjrb52v1h8-systemd-261.2/lib/systemd/systemd-timesyncd
338 systemd+ /nix/store/baxgs06659w4i6ggsfr56srjrb52v1h8-systemd-261.2/lib/systemd/systemd-resolved
347 root /nix/store/baxgs06659w4i6ggsfr56srjrb52v1h8-systemd-261.2/lib/systemd/systemd-udevd
391 root [psimon]
472 systemd+ /nix/store/baxgs06659w4i6ggsfr56srjrb52v1h8-systemd-261.2/lib/systemd/systemd-networkd
490 root [watchdogd]
491 root [kworker/R-cfg80211]
498 root [kworker/R-edac-poller]
569 message+ /nix/store/wqy7b0gz66x4gmkwsyf8dwmmbjp9vdi1-dbus-broker-37/bin/dbus-broker-launch --scope system --audit
571 message+ dbus-broker --log 10 --controller 9 --machine-id f5d8c9d86761459a85062f59346f1e37 --max-bytes 536870912 --max-fds 4096 --max-matches 16384 --audit
575 nscd /nix/store/agzqciyfjb2i3vj3ln9vgfmbgxw4xhzj-nsncd-1.5.2/bin/nsncd
578 root /nix/store/606qg4ms8vkri6cpja5b3p84xm9ximfc-qemu-host-cpu-only-11.1.0-ga/bin/qemu-ga --statedir /run/qemu-ga
580 root sshd: /nix/store/5j1k7bxsljgrgql1cmr98wng8lpk2idn-openssh-10.5p1/bin/sshd -D -f /etc/ssh/sshd_config [listener] 0 of 10-100 startups
592 root /nix/store/zs0ldgp3mgzpxz5y7igqcpj9gpsr8mny-tailscale-1.102.4/bin/tailscaled --state=/var/lib/tailscale/tailscaled.state --socket=/run/tailscale/tailscaled.sock --port=41641 --tun tailscale0 --state=/var/lib/hermes/tailscale/tailscaled.state
593 hhstream /nix/store/pgpmwfkh9xgby9iqv17cf9zhqm7njybw-hyprland-0.56.2/bin/Hyprland --config /nix/store/jbjf3qk6q1jkyhh8x39087dvcsfznx82-hyprland.conf
601 root /nix/store/baxgs06659w4i6ggsfr56srjrb52v1h8-systemd-261.2/lib/systemd/systemd-logind
615 root /nix/store/mqvbf0flamqaq9c3ihb496aahag897n0-util-linux-2.42.3-bin/bin/agetty --login-program /nix/store/wbsjibdqfdcs4gcyx87sb94c4wy3jx05-shadow-4.20.2/bin/login --issue-file /etc/issue:/etc/issue.d:/run/issue:/run/issue.d --noclear --keep-baud tty1 115200,38400,9600 linux
651 hermes /run/current-system/systemd/lib/systemd/systemd --user --deserialize=38
661 hermes (sd-pam)
670 hermes /var/lib/hermes/.hermes/tools/python-3.14.7+20260901-linux-x64/bin/python3 -I -c import os, re, sys os.environ.pop('PYTHONHOME', None) os.environ.pop('PYTHONPATH', None) sys.path.insert(0, '/var/lib/hermes/.hermes/hermes-agent') if sys.argv[1:2] == ['--print-runtime-command']: sys.dont_write_bytecode = True from hermes_constants import get_default_hermes_root os.environ['HERMES_HOME'] = os.environ.get('HERMES_HOME') or str(get_default_hermes_root()) if sys.argv[1:2] == ['--print-runtime-command']: from pathlib import Path from hermes_cli._launchers import print_runtime_command print_runtime_command(Path('/var/lib/hermes/.hermes/hermes-agent'), sys.argv[2:]) sys.exit(0) import hermes_bootstrap if sys.argv[1:2] == ['--run-module']: import runpy if len(sys.argv) < 3: sys.exit('hermes: --run-module needs a module') module = sys.argv.pop(2) del sys.argv[1] runpy.run_module(module, run_name='__main__', alter_sys=True) sys.exit(0) from hermes_cli.main import main sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0]) sys.exit(main()) gateway run
677 hhstream /nix/store/pgpmwfkh9xgby9iqv17cf9zhqm7njybw-hyprland-0.56.2/bin/Hyprland --config /nix/store/jbjf3qk6q1jkyhh8x39087dvcsfznx82-hyprland.conf
698 hhstream /nix/store/rk9kpnq5rd72yhfyw91a8kp5bix1ywfp-swaybg-1.2.2/bin/swaybg -i /nix/store/p6c8877bwk2h7rqq1j33p7rgs2naik9a-agent-eyes.jpg -m fill
724 caddy /nix/store/z7n5vmb63i55f9xzish6fhyxlmwy6mrf-caddy-2.11.4/bin/caddy run --config /etc/caddy/caddy_config --adapter caddyfile
725 hermes /var/lib/hermes/.hermes/tools/python-3.14.7+20260901-linux-x64/bin/python3 -I -c import sys, runpy; sys.path.insert(0, '/var/lib/hermes/.hermes/hermes-agent'); sys.argv = ['/var/lib/hermes/.hermes/hermes-agent/hermes', 'dashboard', '--no-open', '--host', '0.0.0.0', '--skip-build']; runpy.run_path('/var/lib/hermes/.hermes/hermes-agent/hermes', run_name='__main__')
4200 root determinate-nixd --nix-bin /nix/store/cf9fsbghaiv71lpl7jp6v6ha68skrqkb-determinate-nix-3.22.4/bin daemon
4248 root /nix/store/cf9fsbghaiv71lpl7jp6v6ha68skrqkb-determinate-nix-3.22.4/bin/nix-daemon --option json-log-path /nix/var/determinate/determinate-nixd-logger.socket
4254 root /nix/store/myj3yayd9yk0kwbdnf9ixdx3m8x9sm5w-sentry-native-0.13.5/bin/crashpad_handler --no-rate-limit --database=/root/.cache/nix/sentry --metrics-dir=/root/.cache/nix/sentry --url=https://o4506062689927168.ingest.us.sentry.io:443/api/4511237220532224/minidump/?sentry_client=sentry.native/0.13.5&sentry_key=0c7a3f7635bd9f9e92056c3c6e5f0b3a --attachment=/root/.cache/nix/sentry/b5fab905-3ada-49d0-9846-2cab69756924.run/__sentry-event --attachment=/root/.cache/nix/sentry/b5fab905-3ada-49d0-9846-2cab69756924.run/__sentry-breadcrumb1 --attachment=/root/.cache/nix/sentry/b5fab905-3ada-49d0-9846-2cab69756924.run/__sentry-breadcrumb2 --report-id=eb0aa806-28e3-453b-730a-ca337915324b --initial-client-fd=6 --shared-client-connection
6941 hermes /nix/store/wqy7b0gz66x4gmkwsyf8dwmmbjp9vdi1-dbus-broker-37/bin/dbus-broker-launch --scope user
6942 hermes dbus-broker --log 11 --controller 10 --machine-id f5d8c9d86761459a85062f59346f1e37 --max-bytes 100000000000000 --max-fds 25000000000000 --max-matches 5000000000
55730 hermes /nix/store/499dwp4ljzzbx5i5fhlxm6lwzncgqz61-tmux-3.7c/bin/tmux new-session -d -s hhs-pr1-merge -x 200 -y 50 cd /var/lib/hermes/projects/hhstream-keyboard && claude --model opus --effort high "Read CLAUDE-MERGE-BRIEF.md in full and do exactly what it says."; exec bash -i
55731 hermes bash -c cd /var/lib/hermes/projects/hhstream-keyboard && claude --model opus --effort high "Read CLAUDE-MERGE-BRIEF.md in full and do exactly what it says."; exec bash -i
55734 hermes claude --model opus --effort high Read CLAUDE-MERGE-BRIEF.md in full and do exactly what it says.
70277 hermes bash -c cd /var/lib/hermes/projects/pptx-deck/decks/neuralink && claude --model opus --effort high "Read COMMENTS-BRIEF.md in full and do exactly what it says. Do not modify the pptx."; exec bash -i
70280 hermes claude --model opus --effort high Read COMMENTS-BRIEF.md in full and do exactly what it says. Do not modify the pptx.
259625 hermes bash -c cd /var/lib/hermes/worktrees/hhs-mcp && claude --model opus --effort high "Read MCP-BRIEF.md in full and do exactly what it says."; exec bash -i
259628 hermes claude --model opus --effort high Read MCP-BRIEF.md in full and do exactly what it says.
758065 hermes /nix/store/d64q19q1xjdwfhqx6czvrjgrhq0n3lcc-python3-3.14.7/bin/python3.14 /nix/store/4a7ci3sfbkl7zsjpz6r8xsj6lic4z7px-hhstream-mcp-0.1.0/bin/.hhstream-mcp-wrapped
758066 hermes /var/lib/hermes/.hermes/tools/python-3.14.7+20260901-linux-x64/bin/python3 /var/lib/hermes/.hermes/hermes-agent/tools/mcp_death_supervisor.py --parent-pgid 670
807466 root [psimon]
807486 hermes /nix/store/d64q19q1xjdwfhqx6czvrjgrhq0n3lcc-python3-3.14.7/bin/python3.14 /nix/store/wsqlzgwh8f3k6n4pkj3x3ixk25y4rngq-hhstream-0.1.0/bin/.hhstream-pointer-wrapped --socket /run/hhstream/pointer.sock --compositor-runtime-dir /run/hypr-stream --xkb-layout us --xkb-model pc105 --xkb-variant --xkb-options
807487 hermes /nix/store/d64q19q1xjdwfhqx6czvrjgrhq0n3lcc-python3-3.14.7/bin/python3.14 /nix/store/wsqlzgwh8f3k6n4pkj3x3ixk25y4rngq-hhstream-0.1.0/bin/.hhstreamd-wrapped --socket /run/hhstream/control.sock --runtime-dir /run/hhstream --state-dir /var/lib/hermes/hhstream-sessions --config /nix/store/nd8xlpliw4c7s3y3yr2qwbn8056g79yg-hhstream-sessions.toml --compositor-runtime-dir /run/hypr-stream
807511 hhstream /nix/store/d64q19q1xjdwfhqx6czvrjgrhq0n3lcc-python3-3.14.7/bin/python3.14 /nix/store/wsqlzgwh8f3k6n4pkj3x3ixk25y4rngq-hhstream-0.1.0/bin/.hhstream-server-wrapped --socket /run/hypr-stream/stream.sock --fps 30 --format jpeg --capture-command grim --capture-command -c --capture-command -t --capture-command jpeg --capture-command -q --capture-command 80 --capture-command -
831259 root [kworker/0:1-events]
833646 hermes bash -c cd /var/lib/hermes/worktrees/hhs-chat && claude --model opus --effort high "Read COMPANION-CONTEXT.md in full, then stop and wait for my questions — this session is for conversation about the project, not for starting work."; exec bash -i
833649 hermes claude --model opus --effort high Read COMPANION-CONTEXT.md in full, then stop and wait for my questions — this session is for conversation about the project, not for starting work.
847221 root [kworker/u8:3-events_unbound]
863435 root [kworker/u8:2-events_power_efficient]
884175 hermes bash -c cd /var/lib/hermes/worktrees/hhs-capture-speed && claude --model opus --effort high "Read SPIKE-BRIEF.md in full and do exactly what it says."; exec bash -i
884178 hermes claude --model opus --effort high Read SPIKE-BRIEF.md in full and do exactly what it says.
890506 root [kworker/u8:1-flush-253:0]
892374 root [kworker/1:0-events]
898144 hermes /run/current-system/sw/bin/bash -lic set +m; cd /var/lib/hermes/worktrees/hhs-grimrs && env -u XAI_API_KEY grok -p "Read GRIMRS-BRIEF.md in full and do exactly what it says." --cwd /var/lib/hermes/worktrees/hhs-grimrs --max-turns 100 --always-approve --output-format json 2>&1
898162 hermes grok -p Read GRIMRS-BRIEF.md in full and do exactly what it says. --cwd /var/lib/hermes/worktrees/hhs-grimrs --max-turns 100 --always-approve --output-format json
898165 root [kworker/1:1-events]
899058 hermes node /var/lib/hermes/.hermes/lsp/bin/pyright-langserver --stdio
900395 root [kworker/0:0-events]
903719 root [kworker/u8:5-flush-253:0]
906223 root [kworker/1:2-cgroup_release]
911709 root [kworker/0:2]
914615 hermes /nix/store/6nhmbaws8lqd8bvzsx4m0gw36ssyg2iy-foot-1.28.0/bin/foot -H -f monospace:size=20 /nix/store/s4hlzmy3ll1bphdsy8vjdpcw447vbr93-cowsay-3.8.4/bin/cowsay hahaha I can see the world now
915090 hermes /bin/sh /var/lib/hermes/worktrees/hhs-grimrs/spikes/capture-speed/001a-grim-rs/run-cells.sh hh-foot-utjcf2
915093 hermes /run/current-system/sw/bin/bash -O extglob -c snap=$(command cat <&3); builtin shopt -s extglob 2>/dev/null; builtin shopt -s expand_aliases 2>/dev/null; builtin eval -- "$snap"; builtin export GROK_AGENT=1; builtin export PWD="$(builtin pwd)"; unalias find 2>/dev/null || true; find() { local __grok_bin=''; [ -x "$__grok_bin" ] || __grok_bin=$(command -v bfs 2>/dev/null) || __grok_bin=''; if [ -z "$__grok_bin" ]; then command find "$@"; return; fi; if [[ -z ${ZSH_VERSION-} ]] && (( BASH_SUBSHELL > 0 )); then exec -a find "$__grok_bin" "$@"; else (exec -a find "$__grok_bin" "$@"); fi; }; __grok_shadow_find=1; unalias grep 2>/dev/null || true; grep() { local __grok_bin=''; [ -x "$__grok_bin" ] || __grok_bin=$(command -v ugrep 2>/dev/null) || __grok_bin=''; if [ -z "$__grok_bin" ]; then command grep "$@"; return; fi; if [[ -z ${ZSH_VERSION-} ]] && (( BASH_SUBSHELL > 0 )); then exec -a grep "$__grok_bin" -G --ignore-files --hidden -I --exclude-dir=.git --exclude-dir=.svn --exclude-dir=.hg --exclude-dir=.bzr --exclude-dir=.jj --exclude-dir=.sl "$@"; else (exec -a grep "$__grok_bin" -G --ignore-files --hidden -I --exclude-dir=.git --exclude-dir=.svn --exclude-dir=.hg --exclude-dir=.bzr --exclude-dir=.jj --exclude-dir=.sl "$@"); fi; }; __grok_shadow_grep=1; __grok_user_cmd="$1"; builtin declare +x __grok_user_cmd 2>/dev/null; builtin set --; builtin eval "$__grok_user_cmd" 2>&1 -- cd /var/lib/hermes/worktrees/hhs-grimrs && git add \ spikes/capture-speed/001a-grim-rs/shell.nix \ spikes/capture-speed/001a-grim-rs/.gitignore \ spikes/capture-speed/001a-grim-rs/Cargo.toml \ spikes/capture-speed/001a-grim-rs/Cargo.lock \ spikes/capture-speed/001a-grim-rs/src/main.rs \ spikes/capture-speed/001a-grim-rs/run-cells.sh && git status --short && git commit -m "$(cat <<'EOF' spike: grim-rs capture producer for the frozen harness Long-lived zwlr-screencopy client (grim-rs 0.2.0, shm only) that prints the contract JSON. Rust stays in a spike-local shell, not the repo flake. EOF )" && git status --short && git log -1 --oneline
915106 hermes ps -eo pid,user,cmd
@@ -1,13 +0,0 @@
# Spike-local toolchain for route 001a. Rust stays out of the repository flake
# on purpose: this shell is the only place this spike gets rustc and cargo.
#
# Do not set WAYLAND_DISPLAY or XDG_RUNTIME_DIR here. The run commands in
# README.md set WAYLAND_DISPLAY to the absolute socket path. Pointing
# XDG_RUNTIME_DIR at /run/hypr-stream makes libwayland reject the runtime dir.
{ pkgs ? import <nixpkgs> { } }:
pkgs.mkShell {
packages = with pkgs; [
rustc # 1.98 from the machine's nixpkgs
cargo
];
}
@@ -1,859 +0,0 @@
//! Long-lived capture producer for the frozen measurement contract.
//!
//! One Wayland connection (`Grim::new_wlr`, so the crate cannot pick
//! `ext-image-copy-capture-v1`), then warmup frames, then N measured frames.
//! Stdout is exactly one JSON object, on the last line. Everything else goes
//! to stderr.
use std::env;
use std::io::{self, Write};
use std::process::ExitCode;
use std::time::Instant;
use grim_rs::Grim;
const ROUTE: &str = "grim-rs";
const JPEG_QUALITY: u8 = 80;
#[derive(Clone, Copy, PartialEq, Eq, Debug)]
enum Mode {
Raw,
Jpeg,
}
impl Mode {
fn as_str(self) -> &'static str {
match self {
Mode::Raw => "raw",
Mode::Jpeg => "jpeg",
}
}
}
#[derive(Clone, Debug)]
enum Target {
Output(String),
All,
}
impl Target {
fn label(&self) -> &str {
match self {
Target::Output(name) => name,
Target::All => "all",
}
}
}
struct Args {
target: Target,
frames: u32,
warmup: u32,
mode: Mode,
backend: BackendChoice,
}
/// Which constructor to use. `Grim::new()` auto-detects and prefers
/// `ext-image-copy-capture-v1`; the two forcing constructors exist so the
/// table can show what the protocol choice is worth instead of assuming.
#[derive(Clone, Copy, PartialEq, Eq, Debug)]
enum BackendChoice {
Auto,
Wlr,
Ext,
}
impl BackendChoice {
fn open(self) -> Result<(Grim, &'static str), String> {
match self {
BackendChoice::Wlr => Grim::new_wlr()
.map(|grim| (grim, "zwlr-screencopy-v1"))
.map_err(|err| format!("Grim::new_wlr failed: {err}")),
BackendChoice::Ext => Grim::new_ext()
.map(|grim| (grim, "ext-image-copy-capture-v1"))
.map_err(|err| format!("Grim::new_ext failed: {err}")),
// new() prefers ext and falls back to wlr. The crate does not
// report which it chose, so we ask it the same question twice:
// if new_ext() succeeds, new() took ext.
BackendChoice::Auto => {
let probe = Grim::new_ext().is_ok();
Grim::new()
.map(|grim| {
(
grim,
if probe {
"ext-image-copy-capture-v1 (auto)"
} else {
"zwlr-screencopy-v1 (auto, ext unavailable)"
},
)
})
.map_err(|err| format!("Grim::new failed: {err}"))
}
}
}
}
struct Held {
width: u32,
height: u32,
/// JPEG byte length. `None` in raw mode.
jpeg_len: Option<usize>,
/// Pixel or JPEG bytes. Kept alive until the caller drops it, so the
/// frame timer can stop when the bytes are complete and not when they
/// are freed.
bytes: Vec<u8>,
}
fn main() -> ExitCode {
match run() {
Ok(()) => ExitCode::SUCCESS,
Err(err) => {
eprintln!("grimrs-capture: {err}");
ExitCode::from(1)
}
}
}
fn run() -> Result<(), String> {
let args = parse_args(env::args().skip(1))?;
// Absolute WAYLAND_DISPLAY is required. Leaving XDG_RUNTIME_DIR alone is
// also required: this process does not own /run/hypr-stream.
match env::var("WAYLAND_DISPLAY") {
Ok(value) if !value.is_empty() => {}
_ => {
return Err(
"WAYLAND_DISPLAY is empty. Set it to the absolute socket \
path /run/hypr-stream/wayland-1 and do not set \
XDG_RUNTIME_DIR=/run/hypr-stream"
.to_string(),
);
}
}
if env::var("XDG_RUNTIME_DIR").ok().as_deref() == Some("/run/hypr-stream") {
return Err(
"XDG_RUNTIME_DIR=/run/hypr-stream is set. Unset it; libwayland \
refuses a runtime dir the caller does not own"
.to_string(),
);
}
// Neither backend imports a dma-buf: grim-rs 0.2.0 creates a
// NamedTempFile and mmaps it inside every capture call, on both paths
// (see src/platform/wayland/capture.rs). So measuring both is honest on
// a machine with no GPU -- there is no zero-copy path to accidentally
// take, and the protocol choice is the only variable.
let (mut grim, protocol) = args.backend.open()?;
let outputs = grim.get_outputs().map_err(|err| err.to_string())?;
if outputs.is_empty() {
return Err("compositor advertised no outputs".to_string());
}
let layout = describe_outputs(&outputs);
if let Target::Output(name) = &args.target {
if !outputs.iter().any(|output| output.name() == name) {
return Err(format!(
"output {name:?} not found. Outputs: {layout}"
));
}
}
eprintln!(
"grimrs-capture: route {ROUTE} protocol {protocol} \
mode {} target {} warmup {} frames {}",
args.mode.as_str(),
args.target.label(),
args.warmup,
args.frames
);
eprintln!("grimrs-capture: outputs {layout}");
let mut last_warmup_nonzero: Option<u64> = None;
let mut last_warmup_jpeg_ok = true;
for i in 0..args.warmup {
let held = grab(&mut grim, &args.target, args.mode)?;
if args.mode == Mode::Jpeg && !held.bytes.starts_with(&[0xFF, 0xD8]) {
last_warmup_jpeg_ok = false;
}
last_warmup_nonzero = Some(count_nonzero(&held.bytes));
eprintln!(
"grimrs-capture: warmup {}/{} {}x{} bytes {} nonzero {}",
i + 1,
args.warmup,
held.width,
held.height,
held.bytes.len(),
last_warmup_nonzero.unwrap_or(0)
);
}
if args.mode == Mode::Jpeg && !last_warmup_jpeg_ok {
return Err(
"warmup JPEG did not start with FF D8; refusing to measure"
.to_string(),
);
}
let cpu_before = process_cpu_ms()?;
let machine_before = MachineSample::take();
let wall_start = Instant::now();
let mut samples_ms: Vec<f64> = Vec::with_capacity(args.frames as usize);
let mut jpeg_sizes: Vec<u64> = Vec::new();
let mut geometry: Option<(u32, u32)> = None;
for _ in 0..args.frames {
let tick = Instant::now();
let held = grab(&mut grim, &args.target, args.mode)?;
// Touch the length so the buffer is complete before the clock stops.
// Do not read pixels: the contract puts stride and checksum outside
// the measurement.
std::hint::black_box(held.bytes.len());
let ms = tick.elapsed().as_secs_f64() * 1000.0;
match geometry {
None => geometry = Some((held.width, held.height)),
Some((w, h)) if w != held.width || h != held.height => {
return Err(format!(
"geometry changed during the run: {w}x{h} then {}x{}",
held.width, held.height
));
}
Some(_) => {}
}
if let Some(len) = held.jpeg_len {
jpeg_sizes.push(len as u64);
}
samples_ms.push(ms);
// Free between frames, outside the frame timer. wall_ms includes it.
drop(held);
}
let wall_ms = wall_start.elapsed().as_secs_f64() * 1000.0;
let cpu_after = process_cpu_ms()?;
let machine_after = MachineSample::take();
let (width, height) = geometry.ok_or("no measured frames")?;
let per_frame = |delta: Option<f64>| -> Option<f64> {
delta.map(|value| value * 1000.0 / f64::from(args.frames))
};
let system_cpu_ms_per_frame = match (machine_before.system_cpu, machine_after.system_cpu) {
(Some(before), Some(after)) if after >= before => per_frame(Some(after - before)),
_ => None,
};
let compositor_cpu_ms_per_frame =
match (machine_before.compositor_cpu, machine_after.compositor_cpu) {
(Some(before), Some(after)) if after >= before => per_frame(Some(after - before)),
_ => None,
};
let machine_cpu_pct = match (machine_before.system_cpu, machine_after.system_cpu, wall_ms) {
(Some(before), Some(after), wall) if after >= before && wall > 0.0 => {
Some((after - before) * 1000.0 / wall * 100.0)
}
_ => None,
};
let rss_peak = rss_peak_kb();
let cpu_ms_per_frame = match (cpu_before, cpu_after) {
(Some(before), Some(after)) if after >= before => {
Some((after - before) / f64::from(args.frames))
}
_ => None,
};
let opened = match args.backend {
BackendChoice::Auto => "Grim::new() (auto-detect)",
BackendChoice::Wlr => "Grim::new_wlr()",
BackendChoice::Ext => "Grim::new_ext()",
};
let mut notes = format!(
"grim-rs 0.2.0 via {opened}; copy target is a wl_shm buffer \
created inside each capture_* call, not a dma-buf the producer \
imports. Cursor not overlaid (the frozen argv has no cursor flag; \
hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, \
quality {JPEG_QUALITY}, the crate default and the live-stream \
setting. Raw timer stops when capture_* returns, which is after \
the crate copies the mapping into a Vec and swizzles XRGB/ARGB to \
RGBA (and applies any output transform or y-invert). The public \
API does not expose the mapped buffer before that, so raw includes \
work the contract says to leave outside the timer. Percentiles are \
linear interpolation on the sorted samples. cpu_ms_per_frame is \
getrusage(RUSAGE_SELF) utime+stime over the measured loop only, \
divided by frames. Frame buffers are freed between frames; that \
free is inside wall_ms and outside frame_ms. Outputs: {layout}. \
No GPU on this machine; dma-buf was not used and is not claimed. \
This process does not move the pointer and does not start or stop \
a session; the pixels are whatever the compositor is already \
showing."
);
if let Some(nonzero) = last_warmup_nonzero {
notes.push_str(&format!(
" Last warmup frame nonzero bytes: {nonzero}."
));
if nonzero == 0 {
notes.push_str(" WARNING: last warmup frame was all zeros.");
}
}
let report = Report {
route: ROUTE,
language: "rust",
protocol,
output: args.target.label().to_string(),
geometry: [width, height],
mode: args.mode.as_str(),
warmup: args.warmup,
frames: args.frames,
frame_ms: summarize(&samples_ms),
fps_sustained: sustained_fps(&samples_ms),
wall_ms,
cpu_ms_per_frame,
bytes_p50: if args.mode == Mode::Jpeg {
Some(percentile_u64(&jpeg_sizes, 0.50))
} else {
None
},
persistent: true,
rss_peak_mb: rss_peak.map(|kb| (kb as f64 / 1024.0 * 100.0).round() / 100.0),
rss_peak_kb_raw: rss_peak,
system_cpu_ms_per_frame: system_cpu_ms_per_frame.map(|v| (v * 1000.0).round() / 1000.0),
compositor_cpu_ms_per_frame: compositor_cpu_ms_per_frame.map(|v| (v * 1000.0).round() / 1000.0),
machine_cpu_pct: machine_cpu_pct.map(|v| (v * 10.0).round() / 10.0),
frames_bytes_allocated_p50: Some(u64::from(width) * u64::from(height) * 4),
notes,
};
let line = serde_json::to_string(&report.to_json())
.map_err(|err| err.to_string())?;
let mut stdout = io::stdout().lock();
writeln!(stdout, "{line}").map_err(|err| err.to_string())?;
stdout.flush().map_err(|err| err.to_string())?;
Ok(())
}
fn grab(grim: &mut Grim, target: &Target, mode: Mode) -> Result<Held, String> {
let captured = match target {
Target::Output(name) => grim.capture_output(name),
Target::All => grim.capture_all(),
}
.map_err(|err| err.to_string())?;
let width = captured.width();
let height = captured.height();
if width == 0 || height == 0 {
return Err(format!("capture returned {width}x{height}"));
}
match mode {
Mode::Raw => Ok(Held {
width,
height,
jpeg_len: None,
bytes: captured.into_data(),
}),
Mode::Jpeg => {
let jpeg = grim
.to_jpeg_with_quality(captured.data(), width, height, JPEG_QUALITY)
.map_err(|err| err.to_string())?;
let jpeg_len = jpeg.len();
Ok(Held {
width,
height,
jpeg_len: Some(jpeg_len),
bytes: jpeg,
})
}
}
}
fn describe_outputs(outputs: &[grim_rs::Output]) -> String {
let mut ordered: Vec<_> = outputs.iter().collect();
ordered.sort_by(|a, b| {
let ag = a.geometry();
let bg = b.geometry();
ag.x()
.cmp(&bg.x())
.then(ag.y().cmp(&bg.y()))
.then(a.name().cmp(b.name()))
});
ordered
.into_iter()
.map(|output| {
let geometry = output.geometry();
format!(
"{}@{},{} {}x{} scale {}",
output.name(),
geometry.x(),
geometry.y(),
geometry.width(),
geometry.height(),
output.scale()
)
})
.collect::<Vec<_>>()
.join("; ")
}
fn count_nonzero(bytes: &[u8]) -> u64 {
bytes.iter().filter(|byte| **byte != 0).count() as u64
}
fn process_cpu_ms() -> Result<Option<f64>, String> {
let mut usage: libc::rusage = unsafe { std::mem::zeroed() };
let rc = unsafe { libc::getrusage(libc::RUSAGE_SELF, &mut usage) };
if rc != 0 {
return Ok(None);
}
Ok(Some(timeval_ms(usage.ru_utime) + timeval_ms(usage.ru_stime)))
}
fn timeval_ms(value: libc::timeval) -> f64 {
value.tv_sec as f64 * 1000.0 + value.tv_usec as f64 / 1000.0
}
// -- resource counters -------------------------------------------------------
//
// Added by Hermenix after the operator asked for RAM and CPU alongside speed:
// a route that reaches 15 fps by burning a core is not one this machine can
// ship while it is also rendering. Three angles, all sampled around the
// *measured* loop only, so they line up with frame_ms:
//
// * this process (getrusage: CPU and peak RSS)
// * the whole machine (/proc/stat)
// * the compositor (/proc/<Hyprland>/stat)
//
// ru_maxrss is in KiB on Linux despite the name.
fn ticks_per_second() -> f64 {
let value = unsafe { libc::sysconf(libc::_SC_CLK_TCK) };
if value > 0 {
value as f64
} else {
100.0
}
}
fn rss_peak_kb() -> Option<u64> {
let mut usage: libc::rusage = unsafe { std::mem::zeroed() };
let rc = unsafe { libc::getrusage(libc::RUSAGE_SELF, &mut usage) };
if rc != 0 {
return None;
}
Some(usage.ru_maxrss as u64)
}
/// Everything on the machine, all cores, in seconds of CPU.
fn system_cpu_seconds() -> Option<f64> {
let text = std::fs::read_to_string("/proc/stat").ok()?;
let line = text.lines().next()?;
// user nice system idle iowait irq softirq steal ...: sum the busy ones
// (idle and iowait excluded), matching the Python producer's definition.
let values: Vec<f64> = line
.split_whitespace()
.skip(1)
.filter_map(|value| value.parse::<f64>().ok())
.collect();
let busy = values.first()? + values.get(1)? + values.get(2)?
+ values.get(5)? + values.get(6)? + values.get(7)?;
Some(busy / ticks_per_second())
}
/// One process's utime+stime, in seconds.
fn pid_cpu_seconds(pid: u32) -> Option<f64> {
let text = std::fs::read_to_string(format!("/proc/{pid}/stat")).ok()?;
// The comm field is parenthesised and may contain spaces, so split after it.
let after_comm = text.rsplit_once(')')?.1;
let fields: Vec<&str> = after_comm.split_whitespace().collect();
let utime: f64 = fields.get(11)?.parse().ok()?;
let stime: f64 = fields.get(12)?.parse().ok()?;
Some((utime + stime) / ticks_per_second())
}
fn hyprland_pid() -> Option<u32> {
let entries = std::fs::read_dir("/proc").ok()?;
for entry in entries.flatten() {
let name = entry.file_name();
let name = name.to_string_lossy();
let Ok(pid) = name.parse::<u32>() else { continue };
let Ok(comm) = std::fs::read_to_string(format!("/proc/{pid}/comm")) else {
continue;
};
if comm.contains("Hyprland") {
return Some(pid);
}
}
None
}
struct MachineSample {
system_cpu: Option<f64>,
compositor_cpu: Option<f64>,
}
impl MachineSample {
fn take() -> Self {
let pid = hyprland_pid();
MachineSample {
system_cpu: system_cpu_seconds(),
compositor_cpu: pid.and_then(pid_cpu_seconds),
}
}
}
struct Summary {
min: f64,
p50: f64,
mean: f64,
p95: f64,
max: f64,
}
fn summarize(samples: &[f64]) -> Summary {
let mut sorted = samples.to_vec();
sorted.sort_by(|a, b| a.total_cmp(b));
let sum: f64 = sorted.iter().sum();
Summary {
min: sorted[0],
p50: percentile_linear(&sorted, 0.50),
mean: sum / sorted.len() as f64,
p95: percentile_linear(&sorted, 0.95),
max: *sorted.last().expect("at least one sample"),
}
}
fn sustained_fps(samples: &[f64]) -> f64 {
let sum = samples.iter().sum::<f64>();
if sum <= 0.0 {
return 0.0;
}
samples.len() as f64 / (sum / 1000.0)
}
/// Linear interpolation between closest ranks. `samples` must be sorted and
/// non-empty. This is the same definition as numpy's default percentile.
fn percentile_linear(sorted: &[f64], p: f64) -> f64 {
if sorted.len() == 1 {
return sorted[0];
}
let rank = p * (sorted.len() - 1) as f64;
let lo = rank.floor() as usize;
let hi = rank.ceil() as usize;
let frac = rank - lo as f64;
sorted[lo] * (1.0 - frac) + sorted[hi] * frac
}
fn percentile_u64(samples: &[u64], p: f64) -> f64 {
let as_f64: Vec<f64> = samples.iter().map(|value| *value as f64).collect();
let mut sorted = as_f64;
sorted.sort_by(|a, b| a.total_cmp(b));
percentile_linear(&sorted, p)
}
struct Report {
route: &'static str,
language: &'static str,
protocol: &'static str,
output: String,
geometry: [u32; 2],
mode: &'static str,
warmup: u32,
frames: u32,
frame_ms: Summary,
fps_sustained: f64,
wall_ms: f64,
cpu_ms_per_frame: Option<f64>,
bytes_p50: Option<f64>,
persistent: bool,
rss_peak_mb: Option<f64>,
rss_peak_kb_raw: Option<u64>,
system_cpu_ms_per_frame: Option<f64>,
compositor_cpu_ms_per_frame: Option<f64>,
machine_cpu_pct: Option<f64>,
frames_bytes_allocated_p50: Option<u64>,
notes: String,
}
impl Report {
fn to_json(&self) -> serde_json::Value {
serde_json::json!({
"route": self.route,
"language": self.language,
"protocol": self.protocol,
"output": self.output,
"geometry": self.geometry,
"mode": self.mode,
"warmup": self.warmup,
"frames": self.frames,
"frame_ms": {
"min": round3(self.frame_ms.min),
"p50": round3(self.frame_ms.p50),
"mean": round3(self.frame_ms.mean),
"p95": round3(self.frame_ms.p95),
"max": round3(self.frame_ms.max),
},
"fps_sustained": round3(self.fps_sustained),
"wall_ms": round3(self.wall_ms),
"cpu_ms_per_frame": self.cpu_ms_per_frame.map(round3),
"bytes_p50": self.bytes_p50.map(round3),
"persistent": self.persistent,
"rss_peak_mb": self.rss_peak_mb,
"rss_peak_kb_raw": self.rss_peak_kb_raw,
"system_cpu_ms_per_frame": self.system_cpu_ms_per_frame,
"compositor_cpu_ms_per_frame": self.compositor_cpu_ms_per_frame,
"machine_cpu_pct": self.machine_cpu_pct,
"frames_bytes_allocated_p50": self.frames_bytes_allocated_p50,
"notes": self.notes,
})
}
}
fn round3(value: f64) -> f64 {
(value * 1000.0).round() / 1000.0
}
fn parse_args(args: impl IntoIterator<Item = String>) -> Result<Args, String> {
let args: Vec<String> = args.into_iter().collect();
let mut output: Option<String> = None;
let mut all = false;
let mut frames: Option<u32> = None;
let mut warmup: Option<u32> = None;
let mut mode: Option<Mode> = None;
let mut backend = BackendChoice::Auto;
let mut json = false;
let mut index = 0;
while index < args.len() {
let arg = &args[index];
match arg.as_str() {
"--json" => {
if json {
return Err("repeated --json".to_string());
}
json = true;
}
"--all" => {
if all {
return Err("repeated --all".to_string());
}
all = true;
}
"--output" => {
output = Some(need_value(&args, &mut index, "--output")?);
}
"--frames" => {
frames = Some(parse_u32(&need_value(&args, &mut index, "--frames")?, "--frames")?);
}
"--warmup" => {
warmup = Some(parse_u32(&need_value(&args, &mut index, "--warmup")?, "--warmup")?);
}
"--mode" => {
let value = need_value(&args, &mut index, "--mode")?;
mode = Some(match value.as_str() {
"raw" => Mode::Raw,
"jpeg" => Mode::Jpeg,
other => {
return Err(format!(
"--mode must be raw or jpeg, got {other:?}"
))
}
});
}
"--backend" => {
let value = need_value(&args, &mut index, "--backend")?;
backend = match value.as_str() {
"auto" => BackendChoice::Auto,
"wlr" => BackendChoice::Wlr,
"ext" => BackendChoice::Ext,
other => {
return Err(format!(
"--backend must be auto, wlr or ext, got {other:?}"
))
}
};
}
"--help" | "-h" => {
return Err(usage());
}
other => {
return Err(format!("unknown argument {other}. {}", usage()));
}
}
index += 1;
}
if !json {
return Err(format!("--json is required. {}", usage()));
}
if all && output.is_some() {
return Err("--all and --output are mutually exclusive".to_string());
}
let target = if all {
Target::All
} else if let Some(name) = output {
if name.is_empty() {
return Err("--output name is empty".to_string());
}
Target::Output(name)
} else {
return Err(format!("one of --output or --all is required. {}", usage()));
};
let frames = frames.ok_or_else(|| format!("--frames is required. {}", usage()))?;
if frames == 0 {
return Err("--frames must be at least 1".to_string());
}
let warmup = warmup.ok_or_else(|| format!("--warmup is required. {}", usage()))?;
let mode = mode.ok_or_else(|| format!("--mode is required. {}", usage()))?;
Ok(Args {
target,
frames,
warmup,
mode,
backend,
})
}
fn need_value(args: &[String], index: &mut usize, flag: &str) -> Result<String, String> {
let next = index.checked_add(1).and_then(|i| args.get(i));
match next {
Some(value) if !value.starts_with("--") => {
*index += 1;
Ok(value.clone())
}
_ => Err(format!("{flag} needs a value")),
}
}
fn parse_u32(value: &str, flag: &str) -> Result<u32, String> {
value
.parse::<u32>()
.map_err(|_| format!("{flag} must be a non-negative integer, got {value:?}"))
}
fn usage() -> String {
"usage: grimrs-capture (--output NAME | --all) --frames N --warmup W \
--mode raw|jpeg [--backend auto|wlr|ext] --json"
.to_string()
}
#[cfg(test)]
mod tests {
use super::*;
fn parse(line: &str) -> Result<Args, String> {
parse_args(line.split_whitespace().map(str::to_string))
}
#[test]
fn accepts_the_contract_forms() {
let output = parse(
"--output hh-chromium-yrfadn --frames 300 --warmup 10 --mode raw --json",
)
.unwrap();
assert!(matches!(output.target, Target::Output(name) if name == "hh-chromium-yrfadn"));
assert_eq!(output.frames, 300);
assert_eq!(output.warmup, 10);
assert_eq!(output.mode, Mode::Raw);
let all = parse("--all --frames 300 --warmup 10 --mode jpeg --json").unwrap();
assert!(matches!(all.target, Target::All));
assert_eq!(all.mode, Mode::Jpeg);
}
#[test]
fn backend_defaults_to_the_crates_own_choice() {
// The default must be the crate's own auto-detect: forcing a backend
// is an experiment, not the route a production build would take.
let args = parse("--all --frames 1 --warmup 0 --mode raw --json").unwrap();
assert_eq!(args.backend, BackendChoice::Auto);
for (flag, want) in [
("wlr", BackendChoice::Wlr),
("ext", BackendChoice::Ext),
("auto", BackendChoice::Auto),
] {
let line = format!("--all --frames 1 --warmup 0 --mode raw --backend {flag} --json");
assert_eq!(parse(&line).unwrap().backend, want);
}
assert!(parse("--all --frames 1 --warmup 0 --mode raw --backend gpu --json").is_err());
}
#[test]
fn rejects_mixed_targets_and_missing_json() {
assert!(parse("--all --output x --frames 1 --warmup 0 --mode raw --json").is_err());
assert!(parse("--output x --frames 1 --warmup 0 --mode raw").is_err());
assert!(parse("--output x --frames 0 --warmup 0 --mode raw --json").is_err());
assert!(parse("--frames 1 --warmup 0 --mode raw --json").is_err());
assert!(parse("--output x --frames 1 --warmup 0 --mode png --json").is_err());
}
#[test]
fn percentile_matches_linear_interpolation() {
let samples = [1.0, 2.0, 3.0, 4.0];
assert_eq!(percentile_linear(&samples, 0.50), 2.5);
assert_eq!(percentile_linear(&samples, 0.0), 1.0);
assert_eq!(percentile_linear(&samples, 1.0), 4.0);
// rank = 0.95 * 3 = 2.85 -> 3.0 * 0.15 + 4.0 * 0.85
let p95 = percentile_linear(&samples, 0.95);
assert!((p95 - 3.85).abs() < 1e-9);
}
#[test]
fn sustained_fps_uses_the_sum_of_frame_times() {
// Two frames of 100 ms -> 2 / 0.2 = 10 fps.
let fps = sustained_fps(&[100.0, 100.0]);
assert!((fps - 10.0).abs() < 1e-9);
}
#[test]
fn json_object_has_the_contract_keys() {
let report = Report {
route: ROUTE,
language: "rust",
protocol: "zwlr-screencopy-v1",
output: "hh-chromium-yrfadn".to_string(),
geometry: [1280, 720],
mode: "jpeg",
warmup: 10,
frames: 300,
frame_ms: Summary {
min: 1.0,
p50: 2.0,
mean: 3.0,
p95: 4.0,
max: 5.0,
},
fps_sustained: 10.0,
wall_ms: 20.0,
cpu_ms_per_frame: None,
bytes_p50: Some(100.5),
persistent: true,
rss_peak_mb: Some(12.5),
rss_peak_kb_raw: Some(12800),
system_cpu_ms_per_frame: Some(1.0),
compositor_cpu_ms_per_frame: Some(2.0),
machine_cpu_pct: Some(99.9),
frames_bytes_allocated_p50: Some(3686400),
notes: "n".to_string(),
};
let value = report.to_json();
for key in [
"route",
"language",
"protocol",
"output",
"geometry",
"mode",
"warmup",
"frames",
"frame_ms",
"fps_sustained",
"wall_ms",
"cpu_ms_per_frame",
"bytes_p50",
"persistent",
"rss_peak_mb",
"rss_peak_kb_raw",
"system_cpu_ms_per_frame",
"compositor_cpu_ms_per_frame",
"machine_cpu_pct",
"frames_bytes_allocated_p50",
"notes",
] {
assert!(value.get(key).is_some(), "missing {key}");
}
assert!(value["cpu_ms_per_frame"].is_null());
assert_eq!(value["bytes_p50"], 100.5);
assert_eq!(value["persistent"], true);
assert_eq!(value["geometry"][0], 1280);
}
}
Binary file not shown.
-57
View File
@@ -1,57 +0,0 @@
# What crashes the compositor, and exactly when
Recorded from the delegated Claude session that spent its run on this, before that
session was shut down. Its finding is sharper than the one the benchmark itself
produced, so it is written down here rather than left in a tmux transcript.
## The trigger
Reproduced deliberately, not just observed:
1. Hold a live capture session on a **session's headless output** (route 001b's
`--protocol ext` producer, i.e. `ext-image-copy-capture-v1`).
2. Run `hhstream session kill` on that session, which destroys its output.
3. Call `create_frame` on the now-dangling capture session.
→ `SIGSEGV`, core dumped.
The kill is **not** the crash: the compositor dies inside
`ext-image-copy-capture`'s frame creation, on an output that no longer exists.
## The stack
```
#0 CScreenshareFrame::transform() const
#1 CImageCopyCaptureFrame::CImageCopyCaptureFrame(...)
#2 CExtImageCopyCaptureManagerV1Options... (createFrame)
#3 _CExtImageCopyCaptureSessionV1CreateFrame
...
#4 ffi_call_unix64
#10 wl_display_run
```
## Why it matters, and what was done about it
- It explains **three** compositor deaths during the benchmark (the two whose
stack reached `CScreenshareFrame::transform()`, and one in `libaquamarine`),
including one that happened mid-run and invalidated cells: the harness kills a
session between passes while a producer may still hold a capture session on it.
So the harness, not the protocol by itself, was pulling the trigger.
- It is the reason the candidate that was measured and **not** adopted is not
adopted: `ext-image-copy-capture-v1` measured no faster than
`zwlr-screencopy-v1` (78.0 vs 77.7 ms; compositor 67.5 vs 67.2 ms) and its
frame-creation path is the one that segfaults the compositor.
The `feat/pywayland-capture` branch therefore keeps the deprecated
`zwlr-screencopy-v1`, with the switch documented as one constant.
- The crash is in Hyprland (0.56.2, software rendering), not in the harness or in
pywayland. Anything that wants to propose it upstream will need this trigger,
because it is the reproducible part: hold a capture session, destroy the
output underneath it, then ask for a frame.
## Not established
- Whether `zwlr-screencopy-v1` survives the same dangling-output sequence. The
in-process backend in `feat/pywayland-capture` does handle the nearest
equivalent (`global_remove` drops the output, and the next `grab()` reconnects
and re-selects) — but that path has not been exercised against a real session
kill, which is the obvious next experiment.
-163
View File
@@ -1,163 +0,0 @@
# capture-speed — can we drop `grim`, and what would it buy?
**Question** (the operator's): which capture route renders images faster for live
streaming the workspaces to the web and for agent eyes — the shipped `grim`
subprocess, the Rust `grim-rs` crate, or a raw `pywayland` client? And what does
each cost in RAM and CPU?
**Answer**: in-process wins, decisively and cheaply — **~1.5× the frame rate,
~2.8× less of our own CPU, and ~42% less CPU *inside the compositor*** — and
neither Rust nor the newer capture protocol is needed to get it. `grim-rs` is no
faster than the subprocess it would replace and costs **746 MB** of RSS to grim's
18 MB.
## Method
One frozen contract (`HARNESS.md`), one runner (`bench.py`), one target per pass.
- **Target**: a Chromium session on a **CSS-animated** page (no JavaScript, no
timers) so the screen genuinely changes; frames verified to differ between
captures before any run. The still half of the contract lives in `results/still/`
and `results/final-still/` (partial — see "What went wrong").
- **Cells**: one session's container at 1280×720, `raw` and `jpeg q80`, **200
measured frames per run, 2 runs each** (the contract asks 300; 200 is stated
here rather than hidden).
- **Baseline**: `grim` exactly as it ships — one process per frame, fork and
encode inside the timed span.
- **Resources**: producer CPU/RSS (`getrusage`), machine CPU (`/proc/stat`), the
**compositor's own** CPU (`/proc/<Hyprland>/stat`), and an idle **control**
sample with nothing capturing, so an animating page's cost is not attributed
to a route.
- **No GPU**: Hyprland renders through software (kms_swrast/llvmpipe). No route
was tuned to a dma-buf path, and none is claimed. Each producer reports which
protocol actually served its frames.
## Head-to-head — 1280×720, moving screen, one session, sessions run sequentially
| route | mode | p50 ms | fps | own CPU ms | machine CPU ms | **compositor CPU ms** | machine % | **RSS MB** |
|---|---|---|---|---|---|---|---|---|
| `grim` (subprocess, ships today) | jpeg | 117.3 | 8.1 | 5.03 | 247.9 | **116.4** | 199.1 | **18.5** |
| `grim` | raw | 121.4 | 7.9 | 3.86 | 255.2 | 118.4 | 199.4 | 20.5 |
| `grim-rs` (Rust crate) | jpeg | 124.6 | 7.5 | 6.60 | 266.7 | 123.1 | 199.2 | **746.3** |
| `grim-rs` | raw | 120.1 | 7.8 | 1.59 | 254.1 | 119.2 | 198.9 | 746.1 |
| **`pywayland` screencopy** | jpeg | **78.0** | **13.9** | **1.78** | **143.0** | **67.5** | 199.1 | 29.3 |
| `pywayland` screencopy | raw | 76.7 | 14.2 | 0.33 | 140.4 | 67.1 | 198.7 | 24.4 |
| `pywayland` ext-image-copy-capture | jpeg | 77.7 | 14.0 | 1.79 | 142.2 | 67.2 | 199.3 | 29.1 |
| `pywayland` ext-image-copy-capture | raw | 73.5 | 14.4 | 0.31 | 138.5 | 66.2 | 198.9 | 24.5 |
Idle control — nothing capturing at all: **199.1% of one core**, the compositor
alone at 94.1%. Frames are 105–130 KB as JPEG, 2.76 MB as raw PPM.
### What the table says
1. **In-process is ~1.5× faster and much cheaper**: 78 ms vs 117 ms, 13.9 fps vs
8.1, our own CPU 1.8 ms vs 5.0 ms per frame.
2. **The win that matters is the compositor's, not ours**: 67 ms vs 116 ms of
compositor CPU *per frame*. On a 2-core box that is the difference between the
desktop being starved and merely busy.
3. **`grim-rs` is not a candidate**: slightly slower than the subprocess (124.6
vs 117.3 ms), *higher* own CPU (6.6 vs 5.0 ms), and **746 MB RSS** — 40× grim.
Its own notes explain the speed: it allocates a fresh shm buffer and copies
into a `Vec` per capture. The memory figure is reproducible across runs.
4. **screencopy vs ext-image-copy-capture: no measurable difference** (78.0 vs
77.7 ms; compositor 67.5 vs 67.2 ms). The newer protocol buys nothing here, and
its frame-creation path is where one compositor crash landed — so the
recommendation is the older, deprecated-but-working protocol, which is also
what `grim` uses (apples to apples).
5. **The box is saturated in every cell (≈199% of one core)** — and the idle
control is *also* 199%. Rendering the animating screen, not capturing it, is
what fills this machine. No capture change fixes that; what changes is how much
room the compositor has left.
## Where the time actually goes (the finding that reframes it)
`grim` at 1280×720 JPEG q80 spends ~95% of each frame **waiting** — on the fork,
the connect, the compositor's round trip — not encoding: JPEG q80 of a 1280×720
frame costs **1.7 ms** in-process. So the available win is in *not paying the
round trip per frame*, which is exactly what a persistent in-process client does —
and why screencopy and ext-image-copy-capture land within noise of each other.
## What went wrong (each item cost real time; none is hidden)
- **The compositor crashed three times** during measurement (`.Hyprland-wrapped`
SEGV in `CScreenshareFrame::transform()` reached via
`ext-image-copy-capture`'s `create_frame`, plus twice in `libaquamarine`),
taking the stream/sessions/pointer units and both session outputs with it. When
that happened, cells failed and `hhstream status` hung for minutes. **The ext
stack could not be reproduced in isolation** (5 clean frames afterwards), so
this is recorded as *unstable under sustained capture load*, not as "the modern
protocol is broken".
- **Our own client segfaulted on exit 16 times** (`python3.14 … segfault …
libwayland-client.so`) — pywayland/CFFI teardown, *after* the result printed, so
green runs reported exit 139. Fixed with `os._exit(0)` after flushing. A batch
of these coincided in time with a compositor crash, so the fix may matter
beyond tidiness.
- **Two SIGTERM kills of the benchmark** (exit 143, no error line): a 40-minute
`nix-shell` tree does not belong in an agent background process. Moved to a
transient systemd unit, and the runner made resumable.
- **`grim -o NAME` must precede the trailing `-`**, or grim exits 0 having written
**637 bytes of nothing**.
- **grim's only uncompressed format is `ppm`**; `pbm`/`pnm`/`rgb`/`bmp` are
rejected. Raw is *not* cheaper than JPEG for it — writing 2.76 MB down a pipe
costs what encoding costs.
- **turbojpeg reuses a non-NULL `*jpegBuf`**: freeing it and leaving the pointer
makes the next call write into freed memory (heap-corrupting segfault with no
stack trace near the encoder).
- **`display.dispatch(block=False)` never reads the socket** — it processes only
what is already read, so a wait built on it times out against a healthy
compositor. The read is `dispatch(block=True)`.
- **`ext-image-capture-source-v1` has one manager per source kind**
(`ext_output_…`, `ext_foreign_toplevel_…`). Looking for a generic
`ext_image_capture_source_manager_v1` finds nothing and reads as "the modern
protocol is unavailable" — a wrong conclusion this spike drew once. The
toplevel variant means **single-window capture is available**, which matters
for agent eyes.
- **`--all` cells silently changed geometry** (2560→3840) once a second session
existed, so they are excluded from this table rather than averaged.
- **A `--resume` keyed on file existence skips failed cells too**, because a
failure also writes a file; it now requires a parseable result line.
## Verdict: VALIDATED — with a caveat that changes the shape of the answer
- **Replace `grim` with an in-process `pywayland` client on `zwlr-screencopy`**
for the one-session case (live streaming a workspace, agent eyes). ~1.5×
faster, ~2.8× less own CPU, and — the number that decides it — **the
compositor's per-frame CPU nearly halves**.
- **Do not adopt `grim-rs`**: no speed win, 40× the memory.
- **Do not adopt `ext-image-copy-capture` yet** on this Hyprland build: no
measured gain, and its create-frame path appears in a compositor crash stack.
- **Keep raw/ppm out of the wire path**: it is not cheaper than JPEG anywhere
measured.
- **The caveat**: on this software-rendered 2-core VM the machine is saturated by
*rendering* whatever is on screen, in every configuration including idle. The
frame-rate win is real, but no capture change buys headroom — it changes how
much of a saturated box the compositor keeps.
## What a production version would take
- Replace the subprocess behind `src/hhstream/capture.py`'s `CommandCapture` with
a pywayland screencopy client: ~250 lines, vendored protocol XML (the pattern
already used by `src/hhstream_pointer/_protocol/`), and **no new dependency** —
`pywayland` is already a dependency of the package, so the stdlib-only policy
for the frame socket, CLI and broker is untouched.
- JPEG encoding then needs libjpeg-turbo via ctypes (stdlib `zlib` PNG is the
fallback, at 2–3× the bytes) — that *is* a genuine new dependency for that
module, and the one thing worth arguing about in review.
- Single-window capture comes free via
`ext_foreign_toplevel_image_capture_source_manager_v1`; worth a follow-up spike
once the compositor's stability under sustained capture is understood.
## Files
| File | What it is |
|---|---|
| `HARNESS.md` | the frozen measurement contract (argv, JSON, timing boundary, resource fields) |
| `bench.py` | the runner: every route through every cell, raw JSON kept, table derived from the files, `--resume` |
| `producers/grim_producer.py` | route 000, the baseline |
| `producers/wayland_producer.py` | route 001b, raw `pywayland` (screencopy *and* ext-image-copy-capture) |
| `producers/resources.py` | RAM/CPU sampling shared by the Python producers |
| `001a-grim-rs/` | route 001a, the Rust `grim-rs` producer (Rust toolchain confined to its own `shell.nix`) |
| `run-final.sh`, `run-remaining.sh` | the pass drivers, meant to run in a transient systemd unit |
| `chart.py` | the two charts (`python3 chart.py results/final-motion`) |
| `results/final-motion/` | the table above, its chart, and every run's raw stdout/stderr |
| `shell.nix` | the spike's environment — deliberately not in the repo flake |
-223
View File
@@ -1,223 +0,0 @@
#!/usr/bin/env python3
"""Run every route through the frozen contract and print a head-to-head.
The point of a runner rather than a shell loop: the *same* cells get run for
every route with identical arguments, the raw JSON of every run is kept, and the
table is derived from those files instead of from what anyone remembers seeing.
./bench.py --output hh-<session> --frames 300 --runs 2
./bench.py --output hh-<session> --only grim --cells output-jpeg
Routes are commands, so a route that is not built yet simply fails and is
reported as a failed cell rather than quietly omitted.
"""
from __future__ import annotations
import argparse
import json
import os
import shlex
import subprocess
import sys
import time
HERE = os.path.dirname(os.path.abspath(__file__))
PRODUCERS = os.path.join(HERE, "producers")
RUST = os.path.join(HERE, "001a-grim-rs", "target", "release", "grimrs-capture")
ROUTES = {
# name: (description, command builder)
"grim": (
"grim subprocess per frame (shipped)",
lambda o, cell, frames, warmup: [
sys.executable, os.path.join(PRODUCERS, "grim_producer.py"),
*(["--all"] if cell["all"] else ["--output", o]),
"--frames", str(frames), "--warmup", str(warmup), "--mode", cell["mode"], "--json",
],
),
"py-screencopy": (
"pywayland in-process, zwlr-screencopy-v1",
lambda o, cell, frames, warmup: [
sys.executable, os.path.join(PRODUCERS, "wayland_producer.py"),
*(["--all"] if cell["all"] else ["--output", o]),
"--protocol", "screencopy",
"--frames", str(frames), "--warmup", str(warmup), "--mode", cell["mode"], "--json",
],
),
"py-ext": (
"pywayland in-process, ext-image-copy-capture-v1",
lambda o, cell, frames, warmup: [
sys.executable, os.path.join(PRODUCERS, "wayland_producer.py"),
*(["--all"] if cell["all"] else ["--output", o]),
"--protocol", "ext",
"--frames", str(frames), "--warmup", str(warmup), "--mode", cell["mode"], "--json",
],
),
"grim-rs": (
"grim-rs crate (Rust), long-lived process",
lambda o, cell, frames, warmup: [
RUST,
*(["--all"] if cell["all"] else ["--output", o]),
"--frames", str(frames), "--warmup", str(warmup), "--mode", cell["mode"], "--json",
],
),
}
CELLS = [
{"name": "output-raw", "all": False, "mode": "raw"},
{"name": "output-jpeg", "all": False, "mode": "jpeg"},
{"name": "all-raw", "all": True, "mode": "raw"},
{"name": "all-jpeg", "all": True, "mode": "jpeg"},
]
def idle_control(seconds: float) -> dict:
"""What the machine burns with nothing capturing at all.
Without this the resource columns are unreadable: on a software-rendered
compositor an animating page can saturate both cores on its own, and a
per-frame CPU number that includes the animation would be attributed to the
capture route. This is the number to subtract -- and it is also the reason
the verdict has to talk about *what the machine was doing*, not just fps.
"""
sys.path.insert(0, PRODUCERS)
from resources import MachineSample # noqa: E402
before = MachineSample(children=False)
time.sleep(seconds)
after = MachineSample(children=False)
window_ms = seconds * 1000.0
system_ms = None
compositor_ms = None
if before.system_cpu is not None and after.system_cpu is not None:
system_ms = (after.system_cpu - before.system_cpu) * 1000.0
if before.compositor_cpu is not None and after.compositor_cpu is not None:
compositor_ms = (after.compositor_cpu - before.compositor_cpu) * 1000.0
return {
"window_ms": round(window_ms, 1),
"system_cpu_ms": None if system_ms is None else round(system_ms, 1),
"compositor_cpu_ms": None if compositor_ms is None else round(compositor_ms, 1),
"system_cpu_pct": None if system_ms is None else round(system_ms / window_ms * 100.0, 1),
"compositor_cpu_pct": None if compositor_ms is None else round(compositor_ms / window_ms * 100.0, 1),
}
def main() -> int:
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--output", required=True, help="the session container output to measure")
parser.add_argument("--frames", type=int, default=300)
parser.add_argument("--warmup", type=int, default=10)
parser.add_argument("--runs", type=int, default=2)
parser.add_argument("--only", action="append", help="restrict to these route names")
parser.add_argument("--cells", action="append", help="restrict to these cell names")
parser.add_argument("--results", default=os.path.join(HERE, "..", "results"))
parser.add_argument("--env", action="append", default=[], help="KEY=VALUE for the producers")
parser.add_argument("--control-seconds", type=float, default=0.0,
help="sample machine CPU with no captures for this long, as a control")
parser.add_argument("--resume", action="store_true",
help="skip any cell whose result file already exists (for re-running after a kill)")
args = parser.parse_args()
routes = {k: v for k, v in ROUTES.items() if not args.only or k in args.only}
cells = [c for c in CELLS if not args.cells or c["name"] in args.cells]
os.makedirs(args.results, exist_ok=True)
control = None
if args.control_seconds > 0:
print(f"control: sampling {args.control_seconds:.0f}s with nothing capturing...")
control = idle_control(args.control_seconds)
print(f" idle machine CPU: {control['system_cpu_pct']}% of one core "
f"({control['system_cpu_ms'] / control['window_ms'] * 1000:.0f} ms of CPU per second), "
f"compositor {control['compositor_cpu_pct']}% of one core")
with open(os.path.join(args.results, "control.json"), "w") as handle:
json.dump(control, handle, indent=2)
env = dict(os.environ)
for item in args.env:
key, _, value = item.partition("=")
env[key] = value
records: list[dict] = []
failures: list[str] = []
for route, (description, build) in routes.items():
for cell in cells:
for run in range(1, args.runs + 1):
command = build(args.output, cell, args.frames, args.warmup)
label = f"{route}/{cell['name']}/run{run}"
stem = os.path.join(args.results, f"{route}.{cell['name']}.run{run}")
if args.resume and os.path.exists(stem + ".stdout"):
# A *failed* cell also leaves a stdout file, so existence is
# not the test -- a parseable result line is. Otherwise one
# failure is skipped forever and the table shows a permanent
# gap that looks like it was never attempted.
with open(stem + ".stdout") as handle:
lines = [line for line in handle.read().splitlines() if line.startswith("{")]
if lines:
print(f" {label}: already measured, skipped (--resume)")
try:
records.append({"route": route, "description": description,
"cell": cell["name"], "run": run, "rc": 0,
**json.loads(lines[-1])})
except json.JSONDecodeError:
pass
continue
print(f" {label}: previous attempt left no result, re-running")
started = time.monotonic()
proc = subprocess.run(command, capture_output=True, text=True, env=env)
elapsed = time.monotonic() - started
with open(stem + ".stdout", "w") as handle:
handle.write(proc.stdout)
with open(stem + ".stderr", "w") as handle:
handle.write(proc.stderr)
record = {"route": route, "description": description, "cell": cell["name"],
"run": run, "command": " ".join(shlex.quote(part) for part in command),
"rc": proc.returncode, "elapsed_s": round(elapsed, 1)}
last = [line for line in proc.stdout.splitlines() if line.startswith("{")]
if proc.returncode == 0 and last:
record.update(json.loads(last[-1]))
print(f" {label}: p50 {record['frame_ms']['p50']} ms, "
f"{record['fps_sustained']} fps, cpu {record['cpu_ms_per_frame']} ms")
else:
record["error"] = (proc.stderr or proc.stdout).strip().splitlines()[-1:] or ["no output"]
failures.append(label)
print(f" {label}: FAILED rc={proc.returncode} -- {record['error']}")
records.append(record)
with open(os.path.join(args.results, "results.json"), "w") as handle:
json.dump(records, handle, indent=2)
print("\n=== head-to-head (median of runs; jpeg cells are the ones the wire carries) ===")
print("speed | machine cost | footprint")
print(f"{'route':<15} {'cell':<12} {'p50 ms':>8} {'fps':>6} {'own cpu':>8} {'sys cpu':>8} "
f"{'comp cpu':>9} {'machine%':>9} {'rss MB':>7} {'bytes':>9}")
for route in routes:
for cell in cells:
group = [r for r in records if r.get("route") == route and r.get("cell") == cell["name"]]
good = [r for r in group if "frame_ms" in r]
if not good:
print(f"{route:<15} {cell['name']:<12} {'--':>8} {'--':>6} {'--':>8} {'--':>8} {'--':>9} {'--':>9} {'--':>7} (no successful run)")
continue
def median(key, default=0.0):
values = sorted(r.get(key) or 0 for r in good)
return values[len(values) // 2] if values else default
p50 = sorted(r["frame_ms"]["p50"] for r in good)[len(good) // 2]
print(f"{route:<15} {cell['name']:<12} {p50:>8} "
f"{sorted(r['fps_sustained'] for r in good)[len(good)//2]:>6} "
f"{median('cpu_ms_per_frame'):>8} {median('system_cpu_ms_per_frame'):>8} "
f"{median('compositor_cpu_ms_per_frame'):>9} {median('machine_cpu_pct'):>9} "
f"{median('rss_peak_mb'):>7} {int(median('bytes_p50')):>9}")
if control:
print(f"\ncontrol (nothing capturing): {control['system_cpu_pct']}% of one core, "
f"compositor {control['compositor_cpu_ms']} ms/s")
print(" subtract this from 'sys cpu' before attributing CPU to a route.")
if failures:
print(f"\nfailed cells: {', '.join(failures)}")
print(f"\nraw runs: {args.results}/")
return 0
if __name__ == "__main__":
sys.exit(main())
-140
View File
@@ -1,140 +0,0 @@
#!/usr/bin/env python3
"""Turn results/results.json into the two pictures that decide this.
One chart per question, because they answer different things:
- **speed**: p50 frame time and sustained fps per route and cell -- which route is
quickest;
- **machine cost**: the compositor's CPU per frame per route -- which route leaves
the box able to render, which is the question the frame rate does not answer.
The idle control is drawn as a line on the cost chart so a reader can see how
much of the machine was busy before any capture happened.
nix-shell --run "python3 chart.py" # -> results/head-to-head.png
"""
from __future__ import annotations
import json
import os
import statistics
import sys
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt # noqa: E402
HERE = os.path.dirname(os.path.abspath(__file__))
# Which pass to chart: `python3 chart.py results/final-motion`
RESULTS = os.path.join(HERE, sys.argv[1] if len(sys.argv) > 1 else "results")
LABELS = {
"grim": "grim (subprocess)",
"grim-rs": "grim-rs (Rust)",
"py-screencopy": "pywayland\nscreencopy",
"py-ext": "pywayland\next-copy-capture",
}
ORDER = ["grim", "grim-rs", "py-screencopy", "py-ext"]
CELLS = ["output-jpeg", "output-raw", "all-jpeg", "all-raw"]
def load() -> tuple[list[dict], dict | None]:
with open(os.path.join(RESULTS, "results.json")) as handle:
records = json.load(handle)
control = None
try:
with open(os.path.join(RESULTS, "control.json")) as handle:
control = json.load(handle)
except OSError:
pass
return records, control
def cell_value(records: list[dict], route: str, cell: str, key: str) -> float | None:
values = [
r.get(key)
for r in records
if r.get("route") == route and r.get("cell") == cell and r.get(key) is not None
]
if not values:
return None
return statistics.median(values)
def main() -> int:
records, control = load()
routes = [r for r in ORDER if any(rec.get("route") == r for rec in records)]
if not routes:
print("no successful runs in results.json", file=sys.stderr)
return 1
fig, axes = plt.subplots(1, 2, figsize=(15, 6.5), facecolor="#0f1115")
for axis in axes:
axis.set_facecolor("#0f1115")
for spine in axis.spines.values():
spine.set_color("#475569")
axis.tick_params(colors="#cbd5e1", labelsize=9)
axis.grid(axis="y", color="#1e293b", linewidth=0.8)
axis.set_axisbelow(True)
width = 0.8 / len(routes)
colours = ["#f87171", "#fbbf24", "#60a5fa", "#34d399"]
# -- left: speed ---------------------------------------------------------
axis = axes[0]
for index, route in enumerate(routes):
values, positions = [], []
for cell_index, cell in enumerate(CELLS):
value = cell_value(records, route, cell, "fps_sustained")
if value is None:
continue
values.append(value)
positions.append(cell_index + index * width)
axis.bar(positions, values, width, label=LABELS.get(route, route),
color=colours[index % len(colours)], edgecolor="#0f1115")
axis.set_title("speed — sustained frames per second", color="#e2e8f0", fontsize=12)
axis.set_xticks([i + 0.4 - width / 2 for i in range(len(CELLS))])
axis.set_xticklabels([c.replace("-", "\n") for c in CELLS], fontsize=8)
axis.set_ylabel("fps", color="#cbd5e1")
axis.legend(facecolor="#111827", edgecolor="#334155", labelcolor="#e2e8f0", fontsize=8)
# -- right: what it costs the compositor --------------------------------
axis = axes[1]
for index, route in enumerate(routes):
values, positions = [], []
for cell_index, cell in enumerate(CELLS):
value = cell_value(records, route, cell, "compositor_cpu_ms_per_frame")
if value is None:
continue
values.append(value)
positions.append(cell_index + index * width)
axis.bar(positions, values, width, label=LABELS.get(route, route),
color=colours[index % len(colours)], edgecolor="#0f1115")
axis.set_title("cost — the compositor's own CPU per frame (lower is better)",
color="#e2e8f0", fontsize=12)
axis.set_xticks([i + 0.4 - width / 2 for i in range(len(CELLS))])
axis.set_xticklabels([c.replace("-", "\n") for c in CELLS], fontsize=8)
axis.set_ylabel("ms of compositor CPU per frame", color="#cbd5e1")
if control and control.get("compositor_cpu_ms") is not None:
per_frame_note = (
f"idle control: machine {control['system_cpu_pct']}% of one core "
f"with nothing capturing"
)
axis.text(0.02, 0.95, per_frame_note, transform=axis.transAxes,
color="#94a3b8", fontsize=8, va="top")
axis.legend(facecolor="#111827", edgecolor="#334155", labelcolor="#e2e8f0", fontsize=8)
fig.suptitle("hhstream capture routes — 1280×720 container and 2560×720 stitched, "
"jpeg q80 and raw, moving screen, software rendering, no GPU",
color="#e2e8f0", fontsize=11)
fig.tight_layout(rect=(0, 0, 1, 0.95))
out = os.path.join(RESULTS, "head-to-head.png")
fig.savefig(out, dpi=140, facecolor="#0f1115")
print(out)
return 0
if __name__ == "__main__":
sys.exit(main())
-77
View File
@@ -1,77 +0,0 @@
#!/usr/bin/env python3
"""What does a per-frame copy + swizzle cost? Isolating the grim-rs gap.
The suspicion this answers: a compiled Rust library should not be *slower* than a
subprocess, and pure Python should not be faster than either. The explanation
under test is that the routes are not doing the same work.
grim-rs's public API returns owned pixels -- per capture it does
`mmap.to_vec()` plus an XRGB/ARGB -> RGBA swizzle (its own source, and its own
notes in the results). Our producer hands the mapped buffer straight to the
encoder: no allocation, no copy, no conversion.
So this measures what that extra work costs on this box, on its own, with no
compositor involved: allocate a mapping the size of one 1280x720 frame, then time
(a) copying it out and (b) swizzling it, per frame, over 200 frames.
If the copy+swizzle costs about the gap between our number (~78 ms) and grim-rs's
(~124 ms), then the difference is the work, not the language -- and the honest
conclusion is "a client that reuses its buffer beats one that does not", not
"Python beats Rust".
"""
from __future__ import annotations
import mmap
import os
import statistics
import sys
import time
WIDTH, HEIGHT = 1280, 720
STRIDE = WIDTH * 4
SIZE = STRIDE * HEIGHT
FRAMES = 200
def main() -> int:
memory = mmap.mmap(-1, SIZE)
# Fill with something non-trivial so the copy is not optimised away.
memory.write(b"\x01\x02\x03\x04" * (SIZE // 4))
copies: list[float] = []
swizzles: list[float] = []
both: list[float] = []
for _ in range(FRAMES):
start = time.monotonic_ns()
pixels = bytes(memoryview(memory))
copies.append((time.monotonic_ns() - start) / 1e6)
start = time.monotonic_ns()
swapped = bytearray(len(pixels))
swapped[0::4] = pixels[2::4]
swapped[1::4] = pixels[1::4]
swapped[2::4] = pixels[2::4]
swapped[3::4] = pixels[3::4]
swizzles.append((time.monotonic_ns() - start) / 1e6)
both.append(copies[-1] + swizzles[-1])
del pixels, swapped
print(f"frame: {WIDTH}x{HEIGHT}, {SIZE / 1e6:.2f} MB, {FRAMES} frames")
for label, samples in (("copy only (mmap -> Vec)", copies),
("swizzle only (BGRX -> RGBA)", swizzles),
("copy + swizzle", both)):
print(f" {label:<28} p50 {statistics.median(samples):>6.2f} ms "
f"min {min(samples):>6.2f} max {max(samples):>6.2f}")
print("\nAdd 'copy + swizzle' to the in-process route's number to compare it "
"against a library that returns owned pixels.")
memory.close()
# Flush before `os._exit`: os._exit skips the interpreter's cleanup, so
# buffered stdout is dropped -- the first run of this printed nothing at all.
sys.stdout.flush()
os._exit(0)
if __name__ == "__main__":
raise SystemExit(main())
Whitespace-only changes.
@@ -1,30 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2018 Ilia Bozhinov
# Copyright © 2020 Isaac Freund
# Copyright © 2022 wb9688
# Copyright © 2023 i509VCB
#
# Permission to use, copy, modify, distribute, and sell this
# software and its documentation for any purpose is hereby granted
# without fee, provided that the above copyright notice appear in
# all copies and that both that copyright notice and this permission
# notice appear in supporting documentation, and that the name of
# the copyright holders not be used in advertising or publicity
# pertaining to distribution of the software without specific,
# written prior permission. The copyright holders make no
# representations about the suitability of this software for any
# purpose. It is provided "as is" without express or implied
# warranty.
#
# THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS
# SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
# FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
# SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
# WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
# AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
# ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
# THIS SOFTWARE.
from .ext_foreign_toplevel_handle_v1 import ExtForeignToplevelHandleV1 # noqa: F401
from .ext_foreign_toplevel_list_v1 import ExtForeignToplevelListV1 # noqa: F401
@@ -1,191 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2018 Ilia Bozhinov
# Copyright © 2020 Isaac Freund
# Copyright © 2022 wb9688
# Copyright © 2023 i509VCB
#
# Permission to use, copy, modify, distribute, and sell this
# software and its documentation for any purpose is hereby granted
# without fee, provided that the above copyright notice appear in
# all copies and that both that copyright notice and this permission
# notice appear in supporting documentation, and that the name of
# the copyright holders not be used in advertising or publicity
# pertaining to distribution of the software without specific,
# written prior permission. The copyright holders make no
# representations about the suitability of this software for any
# purpose. It is provided "as is" without express or implied
# warranty.
#
# THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS
# SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
# FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
# SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
# WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
# AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
# ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
# THIS SOFTWARE.
from __future__ import annotations
from pywayland.protocol_core import (
Argument,
ArgumentType,
Global,
Interface,
Proxy,
Resource,
)
class ExtForeignToplevelHandleV1(Interface):
"""A mapped toplevel
A :class:`ExtForeignToplevelHandleV1` object represents a mapped toplevel
window. A single app may have multiple mapped toplevels.
"""
name = "ext_foreign_toplevel_handle_v1"
version = 1
class ExtForeignToplevelHandleV1Proxy(Proxy[ExtForeignToplevelHandleV1]):
interface = ExtForeignToplevelHandleV1
@ExtForeignToplevelHandleV1.request()
def destroy(self) -> None:
"""Destroy the :class:`ExtForeignToplevelHandleV1` object
This request should be used when the client will no longer use the
handle or after the closed event has been received to allow destruction
of the object.
When a handle is destroyed, a new handle may not be created by the
server until the toplevel is unmapped and then remapped. Destroying a
toplevel handle is not recommended unless the client is cleaning up
child objects before destroying the
:class:`~pywayland.protocol.ext_foreign_toplevel_list_v1.ExtForeignToplevelListV1`
object, the toplevel was closed or the toplevel handle will not be used
in the future.
Other protocols which extend the :class:`ExtForeignToplevelHandleV1`
interface should require destructors for extension interfaces be called
before allowing the toplevel handle to be destroyed.
"""
self._marshal(0)
self._destroy()
class ExtForeignToplevelHandleV1Resource(Resource):
interface = ExtForeignToplevelHandleV1
@ExtForeignToplevelHandleV1.event()
def closed(self) -> None:
"""The toplevel has been closed
The server will emit no further events on the
:class:`ExtForeignToplevelHandleV1` after this event. Any requests
received aside from the destroy request must be ignored. Upon receiving
this event, the client should destroy the handle.
Other protocols which extend the :class:`ExtForeignToplevelHandleV1`
interface must also ignore requests other than destructors.
"""
self._post_event(0)
@ExtForeignToplevelHandleV1.event()
def done(self) -> None:
"""All information about the toplevel has been sent
This event is sent after all changes in the toplevel state have been
sent.
This allows changes to the :class:`ExtForeignToplevelHandleV1`
properties to be atomically applied. Other protocols which extend the
:class:`ExtForeignToplevelHandleV1` interface may use this event to
also atomically apply any pending state.
This event must not be sent after the
:func:`ExtForeignToplevelHandleV1.closed()` event.
"""
self._post_event(1)
@ExtForeignToplevelHandleV1.event(
Argument(ArgumentType.String),
)
def title(self, title: str) -> None:
"""Title change
The title of the toplevel has changed.
The configured state must not be applied immediately. See
:func:`ExtForeignToplevelHandleV1.done()` for details.
:param title:
:type title:
`ArgumentType.String`
"""
self._post_event(2, title)
@ExtForeignToplevelHandleV1.event(
Argument(ArgumentType.String),
)
def app_id(self, app_id: str) -> None:
"""App_id change
The app id of the toplevel has changed.
The configured state must not be applied immediately. See
:func:`ExtForeignToplevelHandleV1.done()` for details.
:param app_id:
:type app_id:
`ArgumentType.String`
"""
self._post_event(3, app_id)
@ExtForeignToplevelHandleV1.event(
Argument(ArgumentType.String),
)
def identifier(self, identifier: str) -> None:
"""A stable identifier for a toplevel
This identifier is used to check if two or more toplevel handles belong
to the same toplevel.
The identifier is useful for command line tools or privileged clients
which may need to reference an exact toplevel across processes or
instances of the
:class:`~pywayland.protocol.ext_foreign_toplevel_list_v1.ExtForeignToplevelListV1`
global.
The compositor must only send this event when the handle is created.
The identifier must be unique per toplevel and its handles. Two
different toplevels must not have the same identifier. The identifier
is only valid as long as the toplevel is mapped. If the toplevel is
unmapped the identifier must not be reused. An identifier must not be
reused by the compositor to ensure there are no races when sharing
identifiers between processes.
An identifier is a string that contains up to 32 printable ASCII bytes.
An identifier must not be an empty string. It is recommended that a
compositor includes an opaque generation value in identifiers. How the
generation value is used when generating the identifier is
implementation dependent.
:param identifier:
:type identifier:
`ArgumentType.String`
"""
self._post_event(4, identifier)
class ExtForeignToplevelHandleV1Global(Global):
interface = ExtForeignToplevelHandleV1
ExtForeignToplevelHandleV1._gen_c()
ExtForeignToplevelHandleV1.proxy_class = ExtForeignToplevelHandleV1Proxy
ExtForeignToplevelHandleV1.resource_class = ExtForeignToplevelHandleV1Resource
ExtForeignToplevelHandleV1.global_class = ExtForeignToplevelHandleV1Global
@@ -1,150 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2018 Ilia Bozhinov
# Copyright © 2020 Isaac Freund
# Copyright © 2022 wb9688
# Copyright © 2023 i509VCB
#
# Permission to use, copy, modify, distribute, and sell this
# software and its documentation for any purpose is hereby granted
# without fee, provided that the above copyright notice appear in
# all copies and that both that copyright notice and this permission
# notice appear in supporting documentation, and that the name of
# the copyright holders not be used in advertising or publicity
# pertaining to distribution of the software without specific,
# written prior permission. The copyright holders make no
# representations about the suitability of this software for any
# purpose. It is provided "as is" without express or implied
# warranty.
#
# THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS
# SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
# FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
# SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
# WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
# AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
# ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
# THIS SOFTWARE.
from __future__ import annotations
from pywayland.protocol_core import (
Argument,
ArgumentType,
Global,
Interface,
Proxy,
Resource,
)
from .ext_foreign_toplevel_handle_v1 import ExtForeignToplevelHandleV1
class ExtForeignToplevelListV1(Interface):
"""List toplevels
A toplevel is defined as a surface with a role similar to xdg_toplevel.
XWayland surfaces may be treated like toplevels in this protocol.
After a client binds the :class:`ExtForeignToplevelListV1`, each mapped
toplevel window will be sent using the
:func:`ExtForeignToplevelListV1.toplevel()` event.
Clients which only care about the current state can perform a roundtrip
after binding this global.
For each instance of :class:`ExtForeignToplevelListV1`, the compositor must
create a new
:class:`~pywayland.protocol.ext_foreign_toplevel_list_v1.ExtForeignToplevelHandleV1`
object for each mapped toplevel.
If a compositor implementation sends the
:func:`ExtForeignToplevelListV1.finished()` event after the global is
bound, the compositor must not send any
:func:`ExtForeignToplevelListV1.toplevel()` events.
"""
name = "ext_foreign_toplevel_list_v1"
version = 1
class ExtForeignToplevelListV1Proxy(Proxy[ExtForeignToplevelListV1]):
interface = ExtForeignToplevelListV1
@ExtForeignToplevelListV1.request()
def stop(self) -> None:
"""Stop sending events
This request indicates that the client no longer wishes to receive
events for new toplevels.
The Wayland protocol is asynchronous, meaning the compositor may send
further toplevel events until the stop request is processed. The client
should wait for a :func:`ExtForeignToplevelListV1.finished()` event
before destroying this object.
"""
self._marshal(0)
@ExtForeignToplevelListV1.request()
def destroy(self) -> None:
"""Destroy the :class:`ExtForeignToplevelListV1` object
This request should be called either when the client will no longer use
the :class:`ExtForeignToplevelListV1` or after the finished event has
been received to allow destruction of the object.
If a client wishes to destroy this object it should send a
:func:`ExtForeignToplevelListV1.stop()` request and wait for a
:func:`ExtForeignToplevelListV1.finished()` event, then destroy the
handles and then this object.
"""
self._marshal(1)
self._destroy()
class ExtForeignToplevelListV1Resource(Resource):
interface = ExtForeignToplevelListV1
@ExtForeignToplevelListV1.event(
Argument(ArgumentType.NewId, interface=ExtForeignToplevelHandleV1),
)
def toplevel(self, toplevel: ExtForeignToplevelHandleV1) -> None:
"""A toplevel has been created
This event is emitted whenever a new toplevel window is created. It is
emitted for all toplevels, regardless of the app that has created them.
All initial properties of the toplevel (identifier, title, app_id) will
be sent immediately after this event using the corresponding events for
:class:`~pywayland.protocol.ext_foreign_toplevel_list_v1.ExtForeignToplevelHandleV1`.
The compositor will use the :func:`ExtForeignToplevelHandleV1.done()
<pywayland.protocol.ext_foreign_toplevel_list_v1.ExtForeignToplevelHandleV1.done>`
event to indicate when all data has been sent.
:param toplevel:
:type toplevel:
:class:`~pywayland.protocol.ext_foreign_toplevel_list_v1.ExtForeignToplevelHandleV1`
"""
self._post_event(0, toplevel)
@ExtForeignToplevelListV1.event()
def finished(self) -> None:
"""The compositor has finished with the toplevel manager
This event indicates that the compositor is done sending events to this
object. The client should destroy the object. See
:func:`ExtForeignToplevelListV1.destroy()` for more information.
The compositor must not send any more toplevel events after this event.
"""
self._post_event(1)
class ExtForeignToplevelListV1Global(Global):
interface = ExtForeignToplevelListV1
ExtForeignToplevelListV1._gen_c()
ExtForeignToplevelListV1.proxy_class = ExtForeignToplevelListV1Proxy
ExtForeignToplevelListV1.resource_class = ExtForeignToplevelListV1Resource
ExtForeignToplevelListV1.global_class = ExtForeignToplevelListV1Global
@@ -1,27 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2022 Andri Yngvason
# Copyright © 2024 Simon Ser
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
from .ext_foreign_toplevel_image_capture_source_manager_v1 import ExtForeignToplevelImageCaptureSourceManagerV1 # noqa: F401
from .ext_image_capture_source_v1 import ExtImageCaptureSourceV1 # noqa: F401
from .ext_output_image_capture_source_manager_v1 import ExtOutputImageCaptureSourceManagerV1 # noqa: F401
@@ -1,97 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2022 Andri Yngvason
# Copyright © 2024 Simon Ser
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
from __future__ import annotations
from pywayland.protocol_core import (
Argument,
ArgumentType,
Global,
Interface,
Proxy,
Resource,
)
from ..ext_foreign_toplevel_list_v1 import ExtForeignToplevelHandleV1
from .ext_image_capture_source_v1 import ExtImageCaptureSourceV1
class ExtForeignToplevelImageCaptureSourceManagerV1(Interface):
"""Image capture source manager for foreign toplevels
A manager for creating image capture source objects for
:class:`~pywayland.protocol.ext_foreign_toplevel_list_v1.ExtForeignToplevelHandleV1`
objects.
"""
name = "ext_foreign_toplevel_image_capture_source_manager_v1"
version = 1
class ExtForeignToplevelImageCaptureSourceManagerV1Proxy(Proxy[ExtForeignToplevelImageCaptureSourceManagerV1]):
interface = ExtForeignToplevelImageCaptureSourceManagerV1
@ExtForeignToplevelImageCaptureSourceManagerV1.request(
Argument(ArgumentType.NewId, interface=ExtImageCaptureSourceV1),
Argument(ArgumentType.Object, interface=ExtForeignToplevelHandleV1),
)
def create_source(self, toplevel_handle: ExtForeignToplevelHandleV1) -> Proxy[ExtImageCaptureSourceV1]:
"""Create source object for foreign toplevel
Creates a source object for a foreign toplevel handle. Images captured
from this source will show the same content as the toplevel.
:param toplevel_handle:
:type toplevel_handle:
:class:`~pywayland.protocol.ext_foreign_toplevel_list_v1.ExtForeignToplevelHandleV1`
:returns:
:class:`~pywayland.protocol.ext_image_capture_source_v1.ExtImageCaptureSourceV1`
"""
source = self._marshal_constructor(0, ExtImageCaptureSourceV1, toplevel_handle)
return source
@ExtForeignToplevelImageCaptureSourceManagerV1.request()
def destroy(self) -> None:
"""Delete this object
Destroys the manager. This request may be sent at any time by the
client and objects created by the manager will remain valid after its
destruction.
"""
self._marshal(1)
self._destroy()
class ExtForeignToplevelImageCaptureSourceManagerV1Resource(Resource):
interface = ExtForeignToplevelImageCaptureSourceManagerV1
class ExtForeignToplevelImageCaptureSourceManagerV1Global(Global):
interface = ExtForeignToplevelImageCaptureSourceManagerV1
ExtForeignToplevelImageCaptureSourceManagerV1._gen_c()
ExtForeignToplevelImageCaptureSourceManagerV1.proxy_class = ExtForeignToplevelImageCaptureSourceManagerV1Proxy
ExtForeignToplevelImageCaptureSourceManagerV1.resource_class = ExtForeignToplevelImageCaptureSourceManagerV1Resource
ExtForeignToplevelImageCaptureSourceManagerV1.global_class = ExtForeignToplevelImageCaptureSourceManagerV1Global
@@ -1,71 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2022 Andri Yngvason
# Copyright © 2024 Simon Ser
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
from __future__ import annotations
from pywayland.protocol_core import Global, Interface, Proxy, Resource
class ExtImageCaptureSourceV1(Interface):
"""Opaque image capture source object
The image capture source object is an opaque descriptor for a capturable
resource. This resource may be any sort of entity from which an image may
be derived.
Note, because :class:`ExtImageCaptureSourceV1` objects are created from
multiple independent factory interfaces, the
:class:`ExtImageCaptureSourceV1` interface is frozen at version 1.
"""
name = "ext_image_capture_source_v1"
version = 1
class ExtImageCaptureSourceV1Proxy(Proxy[ExtImageCaptureSourceV1]):
interface = ExtImageCaptureSourceV1
@ExtImageCaptureSourceV1.request()
def destroy(self) -> None:
"""Delete this object
Destroys the image capture source. This request may be sent at any time
by the client.
"""
self._marshal(0)
self._destroy()
class ExtImageCaptureSourceV1Resource(Resource):
interface = ExtImageCaptureSourceV1
class ExtImageCaptureSourceV1Global(Global):
interface = ExtImageCaptureSourceV1
ExtImageCaptureSourceV1._gen_c()
ExtImageCaptureSourceV1.proxy_class = ExtImageCaptureSourceV1Proxy
ExtImageCaptureSourceV1.resource_class = ExtImageCaptureSourceV1Resource
ExtImageCaptureSourceV1.global_class = ExtImageCaptureSourceV1Global
@@ -1,98 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2022 Andri Yngvason
# Copyright © 2024 Simon Ser
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
from __future__ import annotations
from pywayland.protocol_core import (
Argument,
ArgumentType,
Global,
Interface,
Proxy,
Resource,
)
from ..wayland import WlOutput
from .ext_image_capture_source_v1 import ExtImageCaptureSourceV1
class ExtOutputImageCaptureSourceManagerV1(Interface):
"""Image capture source manager for outputs
A manager for creating image capture source objects for
:class:`~pywayland.protocol.wayland.WlOutput` objects.
"""
name = "ext_output_image_capture_source_manager_v1"
version = 1
class ExtOutputImageCaptureSourceManagerV1Proxy(Proxy[ExtOutputImageCaptureSourceManagerV1]):
interface = ExtOutputImageCaptureSourceManagerV1
@ExtOutputImageCaptureSourceManagerV1.request(
Argument(ArgumentType.NewId, interface=ExtImageCaptureSourceV1),
Argument(ArgumentType.Object, interface=WlOutput),
)
def create_source(self, output: WlOutput) -> Proxy[ExtImageCaptureSourceV1]:
"""Create source object for output
Creates a source object for an output. Images captured from this source
will show the same content as the output. Some elements may be omitted,
such as cursors and overlays that have been marked as transparent to
capturing.
:param output:
:type output:
:class:`~pywayland.protocol.wayland.WlOutput`
:returns:
:class:`~pywayland.protocol.ext_image_capture_source_v1.ExtImageCaptureSourceV1`
"""
source = self._marshal_constructor(0, ExtImageCaptureSourceV1, output)
return source
@ExtOutputImageCaptureSourceManagerV1.request()
def destroy(self) -> None:
"""Delete this object
Destroys the manager. This request may be sent at any time by the
client and objects created by the manager will remain valid after its
destruction.
"""
self._marshal(1)
self._destroy()
class ExtOutputImageCaptureSourceManagerV1Resource(Resource):
interface = ExtOutputImageCaptureSourceManagerV1
class ExtOutputImageCaptureSourceManagerV1Global(Global):
interface = ExtOutputImageCaptureSourceManagerV1
ExtOutputImageCaptureSourceManagerV1._gen_c()
ExtOutputImageCaptureSourceManagerV1.proxy_class = ExtOutputImageCaptureSourceManagerV1Proxy
ExtOutputImageCaptureSourceManagerV1.resource_class = ExtOutputImageCaptureSourceManagerV1Resource
ExtOutputImageCaptureSourceManagerV1.global_class = ExtOutputImageCaptureSourceManagerV1Global
@@ -1,28 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2021-2023 Andri Yngvason
# Copyright © 2024 Simon Ser
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
from .ext_image_copy_capture_cursor_session_v1 import ExtImageCopyCaptureCursorSessionV1 # noqa: F401
from .ext_image_copy_capture_frame_v1 import ExtImageCopyCaptureFrameV1 # noqa: F401
from .ext_image_copy_capture_manager_v1 import ExtImageCopyCaptureManagerV1 # noqa: F401
from .ext_image_copy_capture_session_v1 import ExtImageCopyCaptureSessionV1 # noqa: F401
@@ -1,185 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2021-2023 Andri Yngvason
# Copyright © 2024 Simon Ser
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
from __future__ import annotations
import enum
from pywayland.protocol_core import (
Argument,
ArgumentType,
Global,
Interface,
Proxy,
Resource,
)
from .ext_image_copy_capture_session_v1 import ExtImageCopyCaptureSessionV1
class ExtImageCopyCaptureCursorSessionV1(Interface):
"""Cursor capture session
This object represents a cursor capture session. It extends the base
capture session with cursor-specific metadata.
"""
name = "ext_image_copy_capture_cursor_session_v1"
version = 1
class error(enum.IntEnum):
duplicate_session = 1
class ExtImageCopyCaptureCursorSessionV1Proxy(Proxy[ExtImageCopyCaptureCursorSessionV1]):
interface = ExtImageCopyCaptureCursorSessionV1
@ExtImageCopyCaptureCursorSessionV1.request()
def destroy(self) -> None:
"""Delete this object
Destroys the session. This request can be sent at any time by the
client.
This request doesn't affect
:class:`~pywayland.protocol.ext_image_copy_capture_v1.ExtImageCopyCaptureFrameV1`
objects created by this object.
"""
self._marshal(0)
self._destroy()
@ExtImageCopyCaptureCursorSessionV1.request(
Argument(ArgumentType.NewId, interface=ExtImageCopyCaptureSessionV1),
)
def get_capture_session(self) -> Proxy[ExtImageCopyCaptureSessionV1]:
"""Get image copy capturer session
Gets the image copy capture session for this cursor session.
The session will produce frames of the cursor image. The compositor may
pause the session when the cursor leaves the captured area.
This request must not be sent more than once, or else the
duplicate_session protocol error is raised.
:returns:
:class:`~pywayland.protocol.ext_image_copy_capture_v1.ExtImageCopyCaptureSessionV1`
"""
session = self._marshal_constructor(1, ExtImageCopyCaptureSessionV1)
return session
class ExtImageCopyCaptureCursorSessionV1Resource(Resource):
interface = ExtImageCopyCaptureCursorSessionV1
@ExtImageCopyCaptureCursorSessionV1.event()
def enter(self) -> None:
"""Cursor entered captured area
Sent when a cursor enters the captured area. It shall be generated
before the "position" and "hotspot" events when and only when a cursor
enters the area.
The cursor enters the captured area when the cursor image intersects
with the captured area. Note, this is different from e.g.
:func:`WlPointer.enter() <pywayland.protocol.wayland.WlPointer.enter>`.
"""
self._post_event(0)
@ExtImageCopyCaptureCursorSessionV1.event()
def leave(self) -> None:
"""Cursor left captured area
Sent when a cursor leaves the captured area. No "position" or "hotspot"
event is generated for the cursor until the cursor enters the captured
area again.
"""
self._post_event(1)
@ExtImageCopyCaptureCursorSessionV1.event(
Argument(ArgumentType.Int),
Argument(ArgumentType.Int),
)
def position(self, x: int, y: int) -> None:
"""Position changed
Cursors outside the image capture source do not get captured and no
event will be generated for them.
The given position is the position of the cursor's hotspot and it is
relative to the main buffer's top left corner in transformed buffer
pixel coordinates. The coordinates may be negative or greater than the
main buffer size.
:param x:
position x coordinates
:type x:
`ArgumentType.Int`
:param y:
position y coordinates
:type y:
`ArgumentType.Int`
"""
self._post_event(2, x, y)
@ExtImageCopyCaptureCursorSessionV1.event(
Argument(ArgumentType.Int),
Argument(ArgumentType.Int),
)
def hotspot(self, x: int, y: int) -> None:
"""Hotspot changed
The hotspot describes the offset between the cursor image and the
position of the input device.
The given coordinates are the hotspot's offset from the origin in
buffer coordinates.
Clients should not apply the hotspot immediately: the hotspot becomes
effective when the next :func:`ExtImageCopyCaptureFrameV1.ready()
<pywayland.protocol.ext_image_copy_capture_v1.ExtImageCopyCaptureFrameV1.ready>`
event is received.
Compositors may delay this event until the client captures a new frame.
:param x:
hotspot x coordinates
:type x:
`ArgumentType.Int`
:param y:
hotspot y coordinates
:type y:
`ArgumentType.Int`
"""
self._post_event(3, x, y)
class ExtImageCopyCaptureCursorSessionV1Global(Global):
interface = ExtImageCopyCaptureCursorSessionV1
ExtImageCopyCaptureCursorSessionV1._gen_c()
ExtImageCopyCaptureCursorSessionV1.proxy_class = ExtImageCopyCaptureCursorSessionV1Proxy
ExtImageCopyCaptureCursorSessionV1.resource_class = ExtImageCopyCaptureCursorSessionV1Resource
ExtImageCopyCaptureCursorSessionV1.global_class = ExtImageCopyCaptureCursorSessionV1Global
@@ -1,298 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2021-2023 Andri Yngvason
# Copyright © 2024 Simon Ser
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
from __future__ import annotations
import enum
from pywayland.protocol_core import (
Argument,
ArgumentType,
Global,
Interface,
Proxy,
Resource,
)
from ..wayland import WlBuffer
class ExtImageCopyCaptureFrameV1(Interface):
"""Image capture frame
This object represents an image capture frame.
The client should attach a buffer, damage the buffer, and then send a
capture request.
If the capture is successful, the compositor must send the frame metadata
(transform, damage, presentation_time in any order) followed by the ready
event.
If the capture fails, the compositor must send the failed event.
"""
name = "ext_image_copy_capture_frame_v1"
version = 1
class error(enum.IntEnum):
no_buffer = 1
invalid_buffer_damage = 2
already_captured = 3
class failure_reason(enum.IntEnum):
unknown = 0
buffer_constraints = 1
stopped = 2
class ExtImageCopyCaptureFrameV1Proxy(Proxy[ExtImageCopyCaptureFrameV1]):
interface = ExtImageCopyCaptureFrameV1
@ExtImageCopyCaptureFrameV1.request()
def destroy(self) -> None:
"""Destroy this object
Destroys the frame. This request can be sent at any time by the client.
"""
self._marshal(0)
self._destroy()
@ExtImageCopyCaptureFrameV1.request(
Argument(ArgumentType.Object, interface=WlBuffer),
)
def attach_buffer(self, buffer: WlBuffer) -> None:
"""Attach buffer to session
Attach a buffer to the session.
The :func:`WlBuffer.release()
<pywayland.protocol.wayland.WlBuffer.release>` request is unused.
The new buffer replaces any previously attached buffer.
This request must not be sent after capture, or else the
already_captured protocol error is raised.
:param buffer:
:type buffer:
:class:`~pywayland.protocol.wayland.WlBuffer`
"""
self._marshal(1, buffer)
@ExtImageCopyCaptureFrameV1.request(
Argument(ArgumentType.Int),
Argument(ArgumentType.Int),
Argument(ArgumentType.Int),
Argument(ArgumentType.Int),
)
def damage_buffer(self, x: int, y: int, width: int, height: int) -> None:
"""Damage buffer
Apply damage to the buffer which is to be captured next. This request
may be sent multiple times to describe a region.
The client indicates the accumulated damage since this
:class:`~pywayland.protocol.wayland.WlBuffer` was last captured. During
capture, the compositor will update the buffer with at least the union
of the region passed by the client and the region advertised by
:func:`ExtImageCopyCaptureFrameV1.damage()`.
When a :class:`~pywayland.protocol.wayland.WlBuffer` is captured for
the first time, or when the client doesn't track damage, the client
must damage the whole buffer.
This is for optimisation purposes. The compositor may use this
information to reduce copying.
These coordinates originate from the upper left corner of the buffer.
If x or y are strictly negative, or if width or height are negative or
zero, the invalid_buffer_damage protocol error is raised.
This request must not be sent after capture, or else the
already_captured protocol error is raised.
:param x:
region x coordinate
:type x:
`ArgumentType.Int`
:param y:
region y coordinate
:type y:
`ArgumentType.Int`
:param width:
region width
:type width:
`ArgumentType.Int`
:param height:
region height
:type height:
`ArgumentType.Int`
"""
self._marshal(2, x, y, width, height)
@ExtImageCopyCaptureFrameV1.request()
def capture(self) -> None:
"""Capture a frame
Capture a frame.
Unless this is the first successful captured frame performed in this
session, the compositor may wait an indefinite amount of time for the
source content to change before performing the copy.
This request may only be sent once, or else the already_captured
protocol error is raised. A buffer must be attached before this request
is sent, or else the no_buffer protocol error is raised.
"""
self._marshal(3)
class ExtImageCopyCaptureFrameV1Resource(Resource):
interface = ExtImageCopyCaptureFrameV1
@ExtImageCopyCaptureFrameV1.event(
Argument(ArgumentType.Uint),
)
def transform(self, transform: int) -> None:
"""Buffer transform
This event is sent before the ready event and holds the transform that
the compositor has applied to the buffer contents.
:param transform:
:type transform:
`ArgumentType.Uint`
"""
self._post_event(0, transform)
@ExtImageCopyCaptureFrameV1.event(
Argument(ArgumentType.Int),
Argument(ArgumentType.Int),
Argument(ArgumentType.Int),
Argument(ArgumentType.Int),
)
def damage(self, x: int, y: int, width: int, height: int) -> None:
"""Buffer damaged region
This event is sent before the ready event. It may be generated multiple
times to describe a region.
The first captured frame in a session will always carry full damage.
Subsequent frames' damaged regions describe which parts of the buffer
have changed since the last ready event.
These coordinates originate in the upper left corner of the buffer.
:param x:
damage x coordinate
:type x:
`ArgumentType.Int`
:param y:
damage y coordinate
:type y:
`ArgumentType.Int`
:param width:
damage width
:type width:
`ArgumentType.Int`
:param height:
damage height
:type height:
`ArgumentType.Int`
"""
self._post_event(1, x, y, width, height)
@ExtImageCopyCaptureFrameV1.event(
Argument(ArgumentType.Uint),
Argument(ArgumentType.Uint),
Argument(ArgumentType.Uint),
)
def presentation_time(self, tv_sec_hi: int, tv_sec_lo: int, tv_nsec: int) -> None:
"""Presentation time of the frame
This event indicates the time at which the frame is presented to the
output in system monotonic time. This event is sent before the ready
event.
The timestamp is expressed as tv_sec_hi, tv_sec_lo, tv_nsec triples,
each component being an unsigned 32-bit value. Whole seconds are in
tv_sec which is a 64-bit value combined from tv_sec_hi and tv_sec_lo,
and the additional fractional part in tv_nsec as nanoseconds. Hence,
for valid timestamps tv_nsec must be in [0, 999999999].
:param tv_sec_hi:
high 32 bits of the seconds part of the timestamp
:type tv_sec_hi:
`ArgumentType.Uint`
:param tv_sec_lo:
low 32 bits of the seconds part of the timestamp
:type tv_sec_lo:
`ArgumentType.Uint`
:param tv_nsec:
nanoseconds part of the timestamp
:type tv_nsec:
`ArgumentType.Uint`
"""
self._post_event(2, tv_sec_hi, tv_sec_lo, tv_nsec)
@ExtImageCopyCaptureFrameV1.event()
def ready(self) -> None:
"""Frame is available for reading
Called as soon as the frame is copied, indicating it is available for
reading.
The buffer may be re-used by the client after this event.
After receiving this event, the client must destroy the object.
"""
self._post_event(3)
@ExtImageCopyCaptureFrameV1.event(
Argument(ArgumentType.Uint),
)
def failed(self, reason: int) -> None:
"""Capture failed
This event indicates that the attempted frame copy has failed.
After receiving this event, the client must destroy the object.
:param reason:
:type reason:
`ArgumentType.Uint`
"""
self._post_event(4, reason)
class ExtImageCopyCaptureFrameV1Global(Global):
interface = ExtImageCopyCaptureFrameV1
ExtImageCopyCaptureFrameV1._gen_c()
ExtImageCopyCaptureFrameV1.proxy_class = ExtImageCopyCaptureFrameV1Proxy
ExtImageCopyCaptureFrameV1.resource_class = ExtImageCopyCaptureFrameV1Resource
ExtImageCopyCaptureFrameV1.global_class = ExtImageCopyCaptureFrameV1Global
@@ -1,139 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2021-2023 Andri Yngvason
# Copyright © 2024 Simon Ser
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
from __future__ import annotations
import enum
from pywayland.protocol_core import (
Argument,
ArgumentType,
Global,
Interface,
Proxy,
Resource,
)
from ..ext_image_capture_source_v1 import ExtImageCaptureSourceV1
from ..wayland import WlPointer
from .ext_image_copy_capture_cursor_session_v1 import ExtImageCopyCaptureCursorSessionV1
from .ext_image_copy_capture_session_v1 import ExtImageCopyCaptureSessionV1
class ExtImageCopyCaptureManagerV1(Interface):
"""Manager to inform clients and begin capturing
This object is a manager which offers requests to start capturing from a
source.
"""
name = "ext_image_copy_capture_manager_v1"
version = 1
class error(enum.IntEnum):
invalid_option = 1
class options(enum.IntFlag):
paint_cursors = 1
class ExtImageCopyCaptureManagerV1Proxy(Proxy[ExtImageCopyCaptureManagerV1]):
interface = ExtImageCopyCaptureManagerV1
@ExtImageCopyCaptureManagerV1.request(
Argument(ArgumentType.NewId, interface=ExtImageCopyCaptureSessionV1),
Argument(ArgumentType.Object, interface=ExtImageCaptureSourceV1),
Argument(ArgumentType.Uint),
)
def create_session(self, source: ExtImageCaptureSourceV1, options: int) -> Proxy[ExtImageCopyCaptureSessionV1]:
"""Capture an image capture source
Create a capturing session for an image capture source.
If the paint_cursors option is set, cursors shall be composited onto
the captured frame. The cursor must not be composited onto the frame if
this flag is not set.
If the options bitfield is invalid, the invalid_option protocol error
is sent.
:param source:
:type source:
:class:`~pywayland.protocol.ext_image_capture_source_v1.ExtImageCaptureSourceV1`
:param options:
:type options:
`ArgumentType.Uint`
:returns:
:class:`~pywayland.protocol.ext_image_copy_capture_v1.ExtImageCopyCaptureSessionV1`
"""
session = self._marshal_constructor(0, ExtImageCopyCaptureSessionV1, source, options)
return session
@ExtImageCopyCaptureManagerV1.request(
Argument(ArgumentType.NewId, interface=ExtImageCopyCaptureCursorSessionV1),
Argument(ArgumentType.Object, interface=ExtImageCaptureSourceV1),
Argument(ArgumentType.Object, interface=WlPointer),
)
def create_pointer_cursor_session(self, source: ExtImageCaptureSourceV1, pointer: WlPointer) -> Proxy[ExtImageCopyCaptureCursorSessionV1]:
"""Capture the pointer cursor of an image capture source
Create a cursor capturing session for the pointer of an image capture
source.
:param source:
:type source:
:class:`~pywayland.protocol.ext_image_capture_source_v1.ExtImageCaptureSourceV1`
:param pointer:
:type pointer:
:class:`~pywayland.protocol.wayland.WlPointer`
:returns:
:class:`~pywayland.protocol.ext_image_copy_capture_v1.ExtImageCopyCaptureCursorSessionV1`
"""
session = self._marshal_constructor(1, ExtImageCopyCaptureCursorSessionV1, source, pointer)
return session
@ExtImageCopyCaptureManagerV1.request()
def destroy(self) -> None:
"""Destroy the manager
Destroy the manager object.
Other objects created via this interface are unaffected.
"""
self._marshal(2)
self._destroy()
class ExtImageCopyCaptureManagerV1Resource(Resource):
interface = ExtImageCopyCaptureManagerV1
class ExtImageCopyCaptureManagerV1Global(Global):
interface = ExtImageCopyCaptureManagerV1
ExtImageCopyCaptureManagerV1._gen_c()
ExtImageCopyCaptureManagerV1.proxy_class = ExtImageCopyCaptureManagerV1Proxy
ExtImageCopyCaptureManagerV1.resource_class = ExtImageCopyCaptureManagerV1Resource
ExtImageCopyCaptureManagerV1.global_class = ExtImageCopyCaptureManagerV1Global
@@ -1,230 +0,0 @@
# This file has been autogenerated by the pywayland scanner
# Copyright © 2021-2023 Andri Yngvason
# Copyright © 2024 Simon Ser
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
from __future__ import annotations
import enum
from pywayland.protocol_core import (
Argument,
ArgumentType,
Global,
Interface,
Proxy,
Resource,
)
from .ext_image_copy_capture_frame_v1 import ExtImageCopyCaptureFrameV1
class ExtImageCopyCaptureSessionV1(Interface):
"""Image copy capture session
This object represents an active image copy capture session.
After a capture session is created, buffer constraint events will be
emitted from the compositor to tell the client which buffer types and
formats are supported for reading from the session. The compositor may re-
send buffer constraint events whenever they change.
To advertise buffer constraints, the compositor must send in no particular
order: zero or more shm_format and dmabuf_format events, zero or one
dmabuf_device event, and exactly one buffer_size event. Then the compositor
must send a done event.
When the client has received all the buffer constraints, it can create a
buffer accordingly, attach it to the capture session using the
attach_buffer request, set the buffer damage using the damage_buffer
request and then send the capture request.
"""
name = "ext_image_copy_capture_session_v1"
version = 1
class error(enum.IntEnum):
duplicate_frame = 1
class ExtImageCopyCaptureSessionV1Proxy(Proxy[ExtImageCopyCaptureSessionV1]):
interface = ExtImageCopyCaptureSessionV1
@ExtImageCopyCaptureSessionV1.request(
Argument(ArgumentType.NewId, interface=ExtImageCopyCaptureFrameV1),
)
def create_frame(self) -> Proxy[ExtImageCopyCaptureFrameV1]:
"""Create a frame
Create a capture frame for this session.
At most one frame object can exist for a given session at any time. If
a client sends a create_frame request before a previous frame object
has been destroyed, the duplicate_frame protocol error is raised.
:returns:
:class:`~pywayland.protocol.ext_image_copy_capture_v1.ExtImageCopyCaptureFrameV1`
"""
frame = self._marshal_constructor(0, ExtImageCopyCaptureFrameV1)
return frame
@ExtImageCopyCaptureSessionV1.request()
def destroy(self) -> None:
"""Delete this object
Destroys the session. This request can be sent at any time by the
client.
This request doesn't affect
:class:`~pywayland.protocol.ext_image_copy_capture_v1.ExtImageCopyCaptureFrameV1`
objects created by this object.
"""
self._marshal(1)
self._destroy()
class ExtImageCopyCaptureSessionV1Resource(Resource):
interface = ExtImageCopyCaptureSessionV1
@ExtImageCopyCaptureSessionV1.event(
Argument(ArgumentType.Uint),
Argument(ArgumentType.Uint),
)
def buffer_size(self, width: int, height: int) -> None:
"""Image capture source dimensions
Provides the dimensions of the source image in buffer pixel
coordinates.
The client must attach buffers that match this size.
:param width:
buffer width
:type width:
`ArgumentType.Uint`
:param height:
buffer height
:type height:
`ArgumentType.Uint`
"""
self._post_event(0, width, height)
@ExtImageCopyCaptureSessionV1.event(
Argument(ArgumentType.Uint),
)
def shm_format(self, format: int) -> None:
"""Shm buffer format
Provides the format that must be used for shared-memory buffers.
This event may be emitted multiple times, in which case the client may
choose any given format.
:param format:
shm format
:type format:
`ArgumentType.Uint`
"""
self._post_event(1, format)
@ExtImageCopyCaptureSessionV1.event(
Argument(ArgumentType.Array),
)
def dmabuf_device(self, device: list) -> None:
"""Dma-buf device
This event advertises the device buffers must be allocated on for dma-
buf buffers.
In general the device is a DRM node. The DRM node type (primary vs.
render) is unspecified. Clients must not rely on the compositor sending
a particular node type. Clients cannot check two devices for equality
by comparing the dev_t value.
:param device:
device dev_t value
:type device:
`ArgumentType.Array`
"""
self._post_event(2, device)
@ExtImageCopyCaptureSessionV1.event(
Argument(ArgumentType.Uint),
Argument(ArgumentType.Array),
)
def dmabuf_format(self, format: int, modifiers: list) -> None:
"""Dma-buf format
Provides the format that must be used for dma-buf buffers.
The client may choose any of the modifiers advertised in the array of
64-bit unsigned integers.
This event may be emitted multiple times, in which case the client may
choose any given format.
:param format:
drm format code
:type format:
`ArgumentType.Uint`
:param modifiers:
drm format modifiers
:type modifiers:
`ArgumentType.Array`
"""
self._post_event(3, format, modifiers)
@ExtImageCopyCaptureSessionV1.event()
def done(self) -> None:
"""All constraints have been sent
This event is sent once when all buffer constraint events have been
sent.
The compositor must always end a batch of buffer constraint events with
this event, regardless of whether it sends the initial constraints or
an update.
"""
self._post_event(4)
@ExtImageCopyCaptureSessionV1.event()
def stopped(self) -> None:
"""Session is no longer available
This event indicates that the capture session has stopped and is no
longer available. This can happen in a number of cases, e.g. when the
underlying source is destroyed, if the user decides to end the image
capture, or if an unrecoverable runtime error has occurred.
The client should destroy the session after receiving this event.
"""
self._post_event(5)
class ExtImageCopyCaptureSessionV1Global(Global):
interface = ExtImageCopyCaptureSessionV1
ExtImageCopyCaptureSessionV1._gen_c()
ExtImageCopyCaptureSessionV1.proxy_class = ExtImageCopyCaptureSessionV1Proxy
ExtImageCopyCaptureSessionV1.resource_class = ExtImageCopyCaptureSessionV1Resource
ExtImageCopyCaptureSessionV1.global_class = ExtImageCopyCaptureSessionV1Global
@@ -1,192 +0,0 @@
#!/usr/bin/env python3
"""Route 000 — the baseline: `grim` as a subprocess, once per frame.
This is the shipped path (``hhstream``'s ``CommandCapture``), measured by the
frozen contract in ``HARNESS.md``: spawn, read to EOF, encode inside the timed
span -- because the fork is the thing we are trying to beat, and leaving it out
would flatter the baseline into something nobody runs.
Deliberately dumb: no connection is held, no buffer is reused, nothing is
cached. It answers "what does today cost?", and the answer is expected to be
dominated by process startup plus one full JPEG encode.
./producers/grim_producer.py --output hh-<session> --frames 300 --warmup 10 \
--mode jpeg --json
"""
from __future__ import annotations
import argparse
import json
import os
import resource
import subprocess
import sys
import time
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
from resources import MachineSample, resource_fields # noqa: E402
#: grim's own words: JPEG q80 at 1280x720 measured ~46 ms here, which is the
#: number ``docs/CONTRACT.md`` cites. `-c` asks for the cursor, exactly as the
#: module's `captureCommand` does, so this measures the shipped shape.
CAPTURE = ["grim", "-c"]
def encode_mode_args(mode: str) -> list[str]:
"""JPEG by default (the wire codec); PPM when ``--mode raw`` is asked for.
PPM is uncompressed RGB -- 2,764,816 bytes for 1280x720, header included --
so `raw` here really is "the pixels, no encoder", and it is the honest way
to price grim's encode separately from its capture. (grim rejects pbm/pnm/
rgb/bmp: ppm is its one uncompressed format.)
"""
if mode == "jpeg":
return ["-t", "jpeg", "-q", "80"]
return ["-t", "ppm"]
def grim_argv(output: str | None, mode: str) -> list[str]:
"""Assemble a grim command line.
**Order is load-bearing.** grim's output file is *positional*, and `-` is
stdout, so `-o NAME` has to come **before** that trailing `-`: with them the
other way round grim captions nothing and writes a 637-byte nothing to the
pipe while still exiting 0. (Found the hard way; the live app in
`src/hhstream_live/__main__.py` inserts `-o` before the trailing `-` for
exactly this reason.)
"""
argv = list(CAPTURE) + encode_mode_args(mode)
if output:
argv += ["-o", output]
return argv + ["-"]
def one_frame(argv: list[str]) -> tuple[float, int]:
"""Return (milliseconds, bytes) for a single subprocess capture."""
start = time.monotonic_ns()
proc = subprocess.run(argv, stdout=subprocess.PIPE, stderr=subprocess.DEVNULL)
data = proc.stdout
elapsed = (time.monotonic_ns() - start) / 1e6
if proc.returncode != 0 or not data:
raise RuntimeError(f"grim failed (exit {proc.returncode}); is WAYLAND_DISPLAY set?")
return elapsed, len(data)
def stats(samples: list[float]) -> dict:
ordered = sorted(samples)
def pick(q: float) -> float:
return ordered[min(len(ordered) - 1, int(q * len(ordered)))]
return {
"min": round(ordered[0], 3),
"p50": round(pick(0.50), 3),
"mean": round(sum(ordered) / len(ordered), 3),
"p95": round(pick(0.95), 3),
"max": round(ordered[-1], 3),
}
def cpu_seconds() -> float:
usage = resource.getrusage(resource.RUSAGE_CHILDREN)
return usage.ru_utime + usage.ru_stime
def main() -> int:
parser = argparse.ArgumentParser(description=__doc__)
target = parser.add_mutually_exclusive_group(required=True)
target.add_argument("--output")
target.add_argument("--all", action="store_true")
parser.add_argument("--frames", type=int, default=300)
parser.add_argument("--warmup", type=int, default=10)
parser.add_argument("--mode", choices=["raw", "jpeg"], default="jpeg")
parser.add_argument("--json", action="store_true")
args = parser.parse_args()
argv = grim_argv(args.output, args.mode)
geometry: list[int] | None = None
for _ in range(args.warmup):
one_frame(argv)
# Resource counters are sampled around the *measured* loop only, so they line
# up with the per-frame times. Children, because this route's cost is a
# process per frame.
before = MachineSample(children=True)
samples: list[float] = []
sizes: list[int] = []
wall_start = time.monotonic_ns()
for index in range(args.frames):
elapsed, size = one_frame(argv)
samples.append(elapsed)
sizes.append(size)
if index == 0:
geometry = _geometry_of(output=args.output, mode=args.mode)
wall_ms = (time.monotonic_ns() - wall_start) / 1e6
after = MachineSample(children=True)
result = {
"route": "grim-subprocess",
"language": "python",
"protocol": "zwlr-screencopy-v1 (grim's client)",
"output": args.output or "all",
"geometry": geometry,
"mode": args.mode,
"warmup": args.warmup,
"frames": args.frames,
"frame_ms": stats(samples),
"fps_sustained": round(len(samples) / (sum(samples) / 1000.0), 2),
"wall_ms": round(wall_ms, 2),
"bytes_p50": sorted(sizes)[len(sizes) // 2],
"persistent": False,
"notes": (
"one grim process per frame, fork and encode inside the timed span; "
f"territory: {'--all (every output stitched)' if args.all else '--output'}; "
+ ("jpeg q80" if args.mode == "jpeg" else "ppm, unencoded")
),
}
result.update(
resource_fields(
before,
after,
frames=args.frames,
bytes_allocated=int(
(geometry[0] * 4 * geometry[1]) if geometry and args.mode == "raw" else 0
)
or None,
window_ms=wall_ms,
)
)
if args.json:
print(json.dumps(result))
else:
print(json.dumps(result, indent=2))
return 0
def _geometry_of(output: str | None, mode: str) -> list[int] | None:
"""Read the frame's own header for its size -- never guess it.
For a PNG/JPEG this is the file header; for pbm, the text header. One extra
capture outside the measurement, purely so the report says what it measured.
"""
import struct
try:
proc = subprocess.run(
list(CAPTURE) + ["-t", "png"] + (["-o", output] if output else []) + ["-"],
stdout=subprocess.PIPE,
stderr=subprocess.DEVNULL,
)
data = proc.stdout
if data[:8] == b"\x89PNG\r\n\x1a\n":
width, height = struct.unpack(">II", data[16:24])
return [width, height]
return None
except Exception:
return None
if __name__ == "__main__":
sys.exit(main())
-142
View File
@@ -1,142 +0,0 @@
#!/usr/bin/env python3
"""Resource measurement for the capture routes: RAM, and CPU from three angles.
Speed is not the whole answer: a route that reaches 15 fps by burning a core is
not one this machine can ship while it is also rendering windows. So every route
reports
- **its own** CPU and peak RSS (`getrusage`, `RUSAGE_CHILDREN` for the
subprocess-per-frame baseline, whose cost lives in children);
- **the machine's** CPU from `/proc/stat`, which is the number that catches a
route that is only fast because somebody else did more work;
- **the compositor's** CPU from `/proc/<hyprland>/stat`, which separates "the
capture is expensive" from "the capture made the compositor expensive".
Everything is sampled around the *measured* loop, never the warm-up, so the
numbers line up with the per-frame times in the result.
"""
from __future__ import annotations
import os
import resource
CLK_TCK = os.sysconf("SC_CLK_TCK")
def _hyprland_pid() -> int | None:
"""The compositor's pid, read from /proc rather than by shelling out.
`systemctl show -p MainPID` would be a CLI call, and a benchmark that shells
out to ask its own question is measuring something other than the route.
"""
try:
for entry in os.listdir("/proc"):
if not entry.isdigit():
continue
try:
with open(f"/proc/{entry}/comm") as handle:
comm = handle.read().strip()
except OSError:
continue
if "Hyprland" in comm:
return int(entry)
except OSError:
return None
return None
def _proc_cpu_ticks(pid: int | None) -> float | None:
"""utime+stime of one process, in seconds."""
if pid is None:
return None
try:
with open(f"/proc/{pid}/stat") as handle:
fields = handle.read().rsplit(")", 1)[1].split()
# After the comm field: state is field 3, so utime/stime (14/15 in the
# full line) land at 11 and 12 here.
return (int(fields[11]) + int(fields[12])) / CLK_TCK
except (OSError, IndexError, ValueError):
return None
def _system_cpu_seconds() -> float | None:
"""Total CPU consumed by everything, all cores, in seconds."""
try:
with open("/proc/stat") as handle:
line = handle.readline()
parts = [int(value) for value in line.split()[1:]]
# user nice system idle iowait irq softirq steal guest guest_nice
busy = parts[0] + parts[1] + parts[2] + parts[5] + parts[6] + parts[7]
return busy / CLK_TCK
except (OSError, ValueError, IndexError):
return None
class MachineSample:
"""One instant of the three CPU counters plus the RSS source."""
def __init__(self, children: bool) -> None:
self.children = children
self.pid = _hyprland_pid()
self.process_cpu = self._own_cpu_seconds()
self.system_cpu = _system_cpu_seconds()
self.compositor_cpu = _proc_cpu_ticks(self.pid)
def _own_cpu_seconds(self) -> float:
usage = resource.getrusage(
resource.RUSAGE_CHILDREN if self.children else resource.RUSAGE_SELF
)
return usage.ru_utime + usage.ru_stime
@property
def rss_peak_kb(self) -> int:
usage = resource.getrusage(
resource.RUSAGE_CHILDREN if self.children else resource.RUSAGE_SELF
)
# ru_maxrss is KiB on Linux, despite the name.
return int(usage.ru_maxrss)
def resource_fields(
before: MachineSample,
after: MachineSample,
frames: int,
bytes_allocated: int | None,
window_ms: float,
cores: int | None = None,
) -> dict:
"""The contract's resource block, from two samples around the measured loop."""
if frames <= 0:
frames = 1
notes: list[str] = []
system_delta_ms = None
if before.system_cpu is not None and after.system_cpu is not None:
system_delta_ms = (after.system_cpu - before.system_cpu) * 1000.0
compositor_ms = None
if before.compositor_cpu is not None and after.compositor_cpu is not None:
compositor_ms = (after.compositor_cpu - before.compositor_cpu) * 1000.0
own_ms = (after.process_cpu - before.process_cpu) * 1000.0
if compositor_ms is None:
notes.append("compositor CPU unreadable (/proc/<Hyprland>/stat)")
cores = cores or os.cpu_count() or 1
machine_pct = None
if system_delta_ms is not None and window_ms > 0:
machine_pct = system_delta_ms / window_ms * 100.0
return {
"cpu_ms_per_frame": round(own_ms / frames, 3),
"rss_peak_mb": round(after.rss_peak_kb / 1024, 2),
"rss_peak_kb_raw": after.rss_peak_kb,
"system_cpu_ms_per_frame": None if system_delta_ms is None else round(system_delta_ms / frames, 3),
"compositor_cpu_ms_per_frame": None if compositor_ms is None else round(compositor_ms / frames, 3),
"machine_cpu_pct": None if machine_pct is None else round(machine_pct, 1),
"machine_cpu_pct_of_all_cores": None if machine_pct is None else round(machine_pct / cores, 1),
"frames_bytes_allocated_p50": bytes_allocated,
"resource_notes": "; ".join(notes),
}
@@ -1,603 +0,0 @@
#!/usr/bin/env python3
"""Route 001b — raw `pywayland`: no subprocess, no helper binary.
One long-lived Wayland connection, one `wl_shm` buffer per output kept alive
between frames, and the capture driven directly over **`zwlr-screencopy`**
(what grim uses, and deprecated) or **`ext-image-copy-capture`** (its
replacement, which this compositor advertises alongside a source manager per
source kind). Which one actually ran is reported, never assumed.
Encoding is `turbojpeg` through `ctypes` -- the library grim itself links, so
both routes pay for the same JPEG. Doing it in-process is the point: the
baseline forks a process per frame and this does not.
Implements the frozen contract in ``HARNESS.md``:
./wayland_producer.py --output hh-<session> --frames 300 --warmup 10 \\
--mode jpeg --json
./wayland_producer.py --all --frames 300 --warmup 10 --mode raw --json
Flags beyond the contract, for debugging a route rather than measuring one:
``--protocol {auto,ext,screencopy}``, ``--list-outputs``, ``--wait-damage``
(ask the compositor to wait for new pixels instead of forcing a repaint --
this is the still-screen question, see the README).
Two pywayland facts this file exists to document: a dispatcher receives the
*proxy* as its first argument, and finishing without `display.disconnect()`
segfaults during interpreter teardown -- after the result is printed, so a
green run reads as exit 139.
"""
from __future__ import annotations
import argparse
import ctypes
import json
import mmap
import os
import resource
import select
import sys
import time
import wl.ext_image_capture_source_v1 as eics
import wl.ext_image_copy_capture_v1 as eicc
import wl.wayland as wl
import wl.wlr_screencopy_unstable_v1 as screencopy
from pywayland.client import Display
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
from resources import MachineSample, resource_fields # noqa: E402
# -- turbojpeg, minus the C headers -------------------------------------------
TJPF_BGRX = 3 # XRGB8888/ARGB8888 in little-endian memory is B,G,R,X
TJSAMP_420 = 2
QUALITY = 80 # grim's -q 80, so the two routes compress the same amount
class TurboJPEG:
"""The encoder grim links, called through ctypes.
`libjpeg-turbo` is a spike-local dependency (see shell.nix); the shipped
package would have to declare it, or publish PNG -- which is 2-3x the bytes.
**Declare every argtype.** Without them ctypes passes pointers through a
default signature and `tjCompress2` reads the wrong width/pitch -- which is a
segfault, not a wrong number, and it took the whole process down mid-run
(the result line never printed) the first time this was written.
"""
def __init__(self) -> None:
self.lib = ctypes.CDLL(os.environ.get("TURBOJPEG_PATH", "libturbojpeg.so"))
self.lib.tjInitCompress.argtypes = []
self.lib.tjInitCompress.restype = ctypes.c_void_p
self.lib.tjCompress2.argtypes = [
ctypes.c_void_p, # tjhandle
ctypes.c_void_p, # srcBuf
ctypes.c_int, # width
ctypes.c_int, # pitch
ctypes.c_int, # height
ctypes.c_int, # pixelFormat
ctypes.POINTER(ctypes.c_void_p), # jpegBuf
ctypes.POINTER(ctypes.c_ulong), # jpegSize
ctypes.c_int, # jpegSubsamp
ctypes.c_int, # jpegQual
ctypes.c_int, # flags
]
self.lib.tjCompress2.restype = ctypes.c_int
self.lib.tjFree.argtypes = [ctypes.c_void_p]
self.lib.tjDestroy.argtypes = [ctypes.c_void_p]
self.handle = self.lib.tjInitCompress()
if not self.handle:
raise RuntimeError("tjInitCompress failed")
self._out = ctypes.c_void_p()
self._size = ctypes.c_ulong()
def compress(self, address: int, width: int, height: int, pitch: int) -> tuple[bytes, float]:
start = time.monotonic_ns()
rc = self.lib.tjCompress2(
ctypes.c_void_p(self.handle),
ctypes.c_void_p(address),
ctypes.c_int(width),
ctypes.c_int(pitch),
ctypes.c_int(height),
ctypes.c_int(TJPF_BGRX),
ctypes.byref(self._out),
ctypes.byref(self._size),
ctypes.c_int(TJSAMP_420),
ctypes.c_int(QUALITY),
0,
)
if rc != 0:
raise RuntimeError(f"tjCompress2 failed: {rc}")
data = ctypes.string_at(self._out, self._size.value)
elapsed = (time.monotonic_ns() - start) / 1e6
self.lib.tjFree(self._out)
# **Reset to NULL after freeing.** turbojpeg treats a non-NULL
# `*jpegBuf` as a caller-supplied destination buffer, so leaving the
# freed pointer in place makes the next call write into freed memory:
# a heap-corrupting segfault, several frames later, with no stack trace
# pointing at the encoder. This is the bug that ate the first jpeg run.
self._out = ctypes.c_void_p()
self._size = ctypes.c_ulong()
return data, elapsed
def close(self) -> None:
if self.handle:
self.lib.tjDestroy(ctypes.c_void_p(self.handle))
self.handle = None
# -- shm buffers ---------------------------------------------------------------
class ShmBuffer:
"""One `wl_shm` pool + buffer + mmap, sized once and reused per frame."""
def __init__(self, shm, width: int, height: int, stride: int, fmt: int):
self.width, self.height, self.stride, self.fmt = width, height, stride, fmt
self.size = stride * height
self.fd = os.memfd_create("hhstream-spike", 0)
os.ftruncate(self.fd, self.size)
self.map = mmap.mmap(self.fd, self.size, prot=mmap.PROT_READ | mmap.PROT_WRITE)
self.pool = shm.create_pool(self.fd, self.size)
self.buffer = self.pool.create_buffer(0, width, height, stride, fmt)
self.address = ctypes.addressof(ctypes.c_char.from_buffer(self.map))
def destroy(self) -> None:
for step in (self.buffer.destroy, self.pool.destroy):
try:
step()
except Exception:
pass
try:
self.map.close()
os.close(self.fd)
except Exception:
pass
class OutputSession:
"""One output, one protocol, one reusable buffer (ext) or a frame per grab."""
def __init__(self, client: "Capture", label: str | None, output, fmt: int, wait_damage: bool):
self.client = client
self.label = label or "?"
self.output = output
self.wait_damage = wait_damage
self.buffer: ShmBuffer | None = None
self.session = None
self.ready = False
self.failed = False
self.geometry: tuple[int, int] | None = None
self._session_done = False
self.formats: list[int] = []
self._wanted_fmt = fmt
if client.protocol.startswith("ext"):
self.source = client.ext_source_mgr.create_source(output)
self.session = client.ext_mgr.create_session(self.source, 0)
self.session.dispatcher["buffer_size"] = self._on_buffer_size
self.session.dispatcher["shm_format"] = self._on_shm_format
self.session.dispatcher["done"] = self._on_session_done
self.client.wait(lambda: self._session_done)
# The session advertises its constraints *before* any frame exists:
# a size, then one `shm_format` per format it can deliver, then
# `done`. Only now can a buffer be allocated -- a frame created
# before this point has nothing to attach (`attach_buffer` on None
# raises, which is how this was found).
if self.geometry is None:
raise RuntimeError(f"the session for {self.label} never sent a buffer size")
width, height = self.geometry
fmt = self.formats[0] if self.formats else 1 # XRGB8888
self.buffer = ShmBuffer(client.shm, width, height, width * 4, fmt)
# -- ext-image-copy-capture --------------------------------------------
def _on_buffer_size(self, session, width: int, height: int) -> None:
self.geometry = (width, height)
def _on_shm_format(self, session, fmt: int) -> None:
self.formats.append(fmt)
def _on_session_done(self, session) -> None:
self._session_done = True
def grab_ext(self) -> None:
self.ready = self.failed = False
frame = self.session.create_frame()
frame.dispatcher["ready"] = lambda proxy: setattr(self, "ready", True)
frame.dispatcher["failed"] = lambda proxy: setattr(self, "failed", True)
frame.dispatcher["transform"] = lambda proxy, t: None
frame.dispatcher["damage"] = lambda proxy, *rest: None
frame.dispatcher["presentation_time"] = lambda proxy, *rest: None
frame.attach_buffer(self.buffer.buffer)
frame.damage_buffer(0, 0, self.buffer.width, self.buffer.height)
frame.capture()
self.client.wait(lambda: self.ready or self.failed)
frame.destroy()
# -- zwlr-screencopy ---------------------------------------------------
def grab_screencopy(self) -> None:
self.ready = self.failed = False
frame = self.client.screencopy_mgr.capture_output(1, self.output)
frame.dispatcher["buffer"] = self._on_sc_buffer
frame.dispatcher["flags"] = lambda proxy, flags: None
frame.dispatcher["damage"] = lambda proxy, *rest: None
frame.dispatcher["ready"] = lambda proxy, *rest: setattr(self, "ready", True)
frame.dispatcher["failed"] = lambda proxy: setattr(self, "failed", True)
self._sc_frame = frame
self.client.wait(lambda: self.buffer is not None)
if self.wait_damage:
frame.copy_with_damage(self.buffer.buffer)
else:
frame.copy(self.buffer.buffer)
self.client.wait(lambda: self.ready or self.failed)
frame.destroy()
def _on_sc_buffer(self, frame, fmt: int, width: int, height: int, stride: int) -> None:
if self.buffer is None or (self.buffer.width, self.buffer.height, self.buffer.stride) != (
width,
height,
stride,
):
if self.buffer:
self.buffer.destroy()
self.buffer = ShmBuffer(self.client.shm, width, height, stride, fmt)
self.geometry = (width, height)
class Capture:
def __init__(self, protocol: str, names_wanted: list[str] | None, want_all: bool, wait_damage: bool):
self.display = Display()
self.display.connect()
self.registry = self.display.get_registry()
self.globals: dict[str, tuple[int, int]] = {}
self._output_globals: list[tuple[int, int]] = []
self._output_names: dict[object, str] = {}
self._output_positions: dict[object, tuple[int, int]] = {}
self._output_proxies: list = []
self.shm = None
self.screencopy_mgr = None
self.ext_mgr = None
self.ext_source_mgr = None
self.protocol = None
self.wait_damage = wait_damage
self.sessions: list[OutputSession] = []
self.stage_ms = {"capture": 0.0, "encode": 0.0}
self._canvas: ctypes.Array | None = None
self._canvas_geometry: tuple[int, int] | None = None
self.registry.dispatcher["global"] = self._on_global
self.display.roundtrip()
if "wl_shm" not in self.globals:
raise RuntimeError("no wl_shm: the compositor cannot hand us pixels")
self.protocol = self._pick_protocol(protocol)
shm_name, shm_version = self.globals["wl_shm"]
self.shm = self.registry.bind(shm_name, wl.WlShm, shm_version)
for name, version in self._output_globals:
output = self.registry.bind(name, wl.WlOutput, min(version, 4))
self._output_proxies.append(output) # a GC'd proxy stops delivering events
output.dispatcher["name"] = self._on_output_name
output.dispatcher["geometry"] = self._on_output_geometry
self.display.roundtrip()
# Sorted by position, not by registry order: `--all` stitches these
# left to right, and "the order the compositor announced its outputs"
# is not a promise about where they are on the desktop.
self.outputs: list[tuple[str, object]] = sorted(
(
(self._output_names.get(proxy, f"output-{index}"), proxy)
for index, proxy in enumerate(self._output_proxies)
),
key=lambda pair: (self._output_positions.get(pair[1], (1 << 30, 0)), pair[0]),
)
if not want_all:
wanted = set(names_wanted or [])
chosen = [pair for pair in self.outputs if pair[0] in wanted]
missing = wanted - {label for label, _ in chosen}
if missing:
raise RuntimeError(
f"no such output: {sorted(missing)}; this compositor has "
f"{sorted(label for label, _ in self.outputs)}"
)
self.outputs = chosen
if not self.outputs:
raise RuntimeError("no outputs to capture")
def open_sessions(self) -> None:
"""Start a capture session per target output. Separate from __init__ so
`--list-outputs` can enumerate without setting any capture up."""
if self.sessions:
return
self.sessions = [
OutputSession(self, label, proxy, fmt=0, wait_damage=self.wait_damage)
for label, proxy in self.outputs
]
# -- registry plumbing -------------------------------------------------
def _on_global(self, registry, name: int, interface: str, version: int) -> None:
self.globals[interface] = (name, version)
if interface == "wl_output":
self._output_globals.append((name, version))
def _on_output_name(self, output, name: str) -> None:
self._output_names[output] = name
def _on_output_geometry(self, output, x: int, y: int, physical_width: int,
physical_height: int, subpixel: int, make: str,
model: str, transform: int) -> None:
"""Where this output is in the layout -- the stitch order for `--all`."""
self._output_positions[output] = (x, y)
def _pick_protocol(self, requested: str) -> str:
have_ext = (
"ext_image_copy_capture_manager_v1" in self.globals
and "ext_output_image_capture_source_manager_v1" in self.globals
)
have_scr = "zwlr_screencopy_manager_v1" in self.globals
if requested == "ext" and not have_ext:
raise RuntimeError("--protocol ext asked for, but the compositor does not advertise it")
if requested == "screencopy" and not have_scr:
raise RuntimeError("--protocol screencopy asked for, but the compositor does not advertise it")
use_ext = have_ext if requested == "auto" else requested == "ext"
if not use_ext and not have_scr:
raise RuntimeError("neither capture protocol is available")
if use_ext:
name, version = self.globals["ext_image_copy_capture_manager_v1"]
self.ext_mgr = self.registry.bind(name, eicc.ExtImageCopyCaptureManagerV1, version)
name, version = self.globals["ext_output_image_capture_source_manager_v1"]
self.ext_source_mgr = self.registry.bind(
name, eics.ExtOutputImageCaptureSourceManagerV1, version
)
return "ext-image-copy-capture-v1"
name, version = self.globals["zwlr_screencopy_manager_v1"]
self.screencopy_mgr = self.registry.bind(name, screencopy.ZwlrScreencopyManagerV1, version)
return "zwlr-screencopy-v1"
# -- dispatch ----------------------------------------------------------
def wait(self, done, timeout: float = 10.0) -> None:
"""Pump events until *done()*, or raise. Never blocks forever.
`display.dispatch(block=False)` processes only events **already read**
from the socket -- it never reads. A wait built on it (plus `select`)
therefore times out on a compositor that is answering perfectly, which
cost an hour of this spike. The read is `dispatch(block=True)`, and the
"never hang" guarantee comes from a real alarm rather than from polling.
"""
import signal
def on_alarm(signum, frame): # pragma: no cover - exercised on failure
raise TimeoutError(
"the compositor never became ready within "
f"{timeout:.0f}s (still screen? see --wait-damage)"
)
previous = signal.signal(signal.SIGALRM, on_alarm)
signal.setitimer(signal.ITIMER_REAL, timeout)
try:
self.display.flush()
while not done():
self.display.dispatch(block=True)
finally:
signal.setitimer(signal.ITIMER_REAL, 0)
signal.signal(signal.SIGALRM, previous)
# -- the measured path -------------------------------------------------
def grab(self) -> tuple[int, int]:
"""One frame of every target output; returns the stitched geometry."""
start = time.monotonic_ns()
for session in self.sessions:
if self.protocol.startswith("ext"):
session.grab_ext()
else:
session.grab_screencopy()
if session.failed:
raise RuntimeError(f"capture failed on {session.label}")
self.stage_ms["capture"] += (time.monotonic_ns() - start) / 1e6
width = sum(session.buffer.width for session in self.sessions)
height = max(session.buffer.height for session in self.sessions)
return width, height
def jpeg(self, encoder: TurboJPEG) -> bytes | None:
"""Encode the frame the wire would carry, stitching first if needed.
The first version of this returned ``None`` for `--all` and left the
stitch "to the caller" -- which no caller did, so the `--all --mode
jpeg` cell reported an encode of 0.0 ms and `bytes_p50: null` while
looking like a successful run. That cell is the one `hhstream`
publishes today, so the silent hole was in the only number that
decides anything.
The stitch is charged to the **encode** stage, not to capture: it is
work this route does to turn N output buffers into one image, and it
is exactly the work `grim -c` does inside its own process.
"""
if len(self.sessions) == 1:
buffer = self.sessions[0].buffer
data, elapsed = encoder.compress(
buffer.address, buffer.width, buffer.height, buffer.stride
)
self.stage_ms["encode"] += elapsed
return data
start = time.monotonic_ns()
width, height, stride, address = self._stitch()
stitch_ms = (time.monotonic_ns() - start) / 1e6
data, elapsed = encoder.compress(address, width, height, stride)
self.stage_ms["encode"] += stitch_ms + elapsed
self.stage_ms["stitch"] = self.stage_ms.get("stitch", 0.0) + stitch_ms
return data
def _stitch(self) -> tuple[int, int, int, int]:
"""Copy every output buffer side by side into one canvas.
Allocated once and reused, like the per-output buffers: a fresh
10 MB canvas per frame would be measuring the allocator. Rows are
copied one `memmove` at a time because the canvas stride is wider
than any single output's.
"""
width = sum(session.buffer.width for session in self.sessions)
height = max(session.buffer.height for session in self.sessions)
stride = width * 4
if self._canvas is None or self._canvas_geometry != (width, height):
self._canvas = ctypes.create_string_buffer(stride * height)
self._canvas_geometry = (width, height)
base = ctypes.addressof(self._canvas)
x_offset = 0
for session in self.sessions:
buffer = session.buffer
row_bytes = buffer.width * 4
for row in range(min(buffer.height, height)):
ctypes.memmove(
base + row * stride + x_offset * 4,
buffer.address + row * buffer.stride,
row_bytes,
)
x_offset += buffer.width
return width, height, stride, base
def buffer_bytes(self) -> int | None:
"""Every byte of frame buffer this route holds, canvas included."""
held = [session.buffer.size for session in self.sessions if session.buffer]
if not held:
return None
total = sum(held)
if self._canvas is not None:
total += len(self._canvas)
return total
def close(self) -> None:
for session in getattr(self, "sessions", []):
if session.buffer:
session.buffer.destroy()
self.display.disconnect()
def stats(samples: list[float]) -> dict:
ordered = sorted(samples)
def pick(q: float) -> float:
return ordered[min(len(ordered) - 1, int(q * len(ordered)))]
return {
"min": round(ordered[0], 3),
"p50": round(pick(0.50), 3),
"mean": round(sum(ordered) / len(ordered), 3),
"p95": round(pick(0.95), 3),
"max": round(ordered[-1], 3),
}
def main() -> int:
parser = argparse.ArgumentParser(description=__doc__)
target = parser.add_mutually_exclusive_group(required=True)
target.add_argument("--output", action="append", help="wl_output name; repeat for several")
target.add_argument("--all", action="store_true", help="every output, stitched")
target.add_argument("--list-outputs", action="store_true")
parser.add_argument("--frames", type=int, default=300)
parser.add_argument("--warmup", type=int, default=10)
parser.add_argument("--mode", choices=["raw", "jpeg"], default="jpeg")
parser.add_argument("--protocol", choices=["auto", "ext", "screencopy"], default="auto")
parser.add_argument("--wait-damage", action="store_true", help="wait for new pixels instead of forcing a repaint")
parser.add_argument("--json", action="store_true")
args = parser.parse_args()
if args.list_outputs:
client = Capture("auto", None, True, False)
for label, _ in client.outputs:
print(label)
client.close()
return 0
client = Capture(
args.protocol, args.output, args.all, wait_damage=args.wait_damage
)
client.open_sessions()
encoder = TurboJPEG() if args.mode == "jpeg" else None
geometry: tuple[int, int] | None = None
try:
for _ in range(args.warmup):
client.grab()
if encoder:
client.jpeg(encoder)
# Sampled around the measured loop only: a persistent route's cost is in
# this process, so RUSAGE_SELF, not children.
before = MachineSample(children=False)
client.stage_ms = {"capture": 0.0, "encode": 0.0}
samples: list[float] = []
sizes: list[int] = []
wall_start = time.monotonic_ns()
for _ in range(args.frames):
frame_start = time.monotonic_ns()
geometry = client.grab()
data = client.jpeg(encoder) if encoder else None
samples.append((time.monotonic_ns() - frame_start) / 1e6)
if data is not None:
sizes.append(len(data))
wall_ms = (time.monotonic_ns() - wall_start) / 1e6
after = MachineSample(children=False)
finally:
protocol = client.protocol
buffer_bytes = client.buffer_bytes()
client.close()
if encoder:
encoder.close()
result = {
"route": "pywayland-in-process",
"language": "python",
"protocol": protocol,
"output": args.output[0] if args.output else "all",
"geometry": list(geometry) if geometry else None,
"mode": args.mode,
"warmup": args.warmup,
"frames": len(samples),
"frame_ms": stats(samples),
"stage_ms": {
"capture": round(client.stage_ms["capture"] / max(1, len(samples)), 3),
"encode": round(client.stage_ms["encode"] / max(1, len(samples)), 3),
"stitch": round(client.stage_ms.get("stitch", 0.0) / max(1, len(samples)), 3),
},
"fps_sustained": round(len(samples) / (sum(samples) / 1000.0), 2) if samples else 0.0,
"wall_ms": round(wall_ms, 2),
"bytes_p50": sorted(sizes)[len(sizes) // 2] if sizes else None,
"persistent": True,
"notes": (
f"one connection, buffers reused, turbojpeg q{QUALITY} in-process; "
f"{len(client.outputs)} output(s); "
+ ("waiting for damage" if args.wait_damage else "forcing a repaint per frame")
+ (f"; stitched {len(client.outputs)} outputs" if len(client.outputs) > 1 else "")
),
}
result.update(
resource_fields(
before,
after,
frames=len(samples),
bytes_allocated=buffer_bytes,
window_ms=wall_ms,
)
)
print(json.dumps(result) if args.json else json.dumps(result, indent=2))
# **Exit without running teardown.** pywayland/CFFI finalizers segfault inside
# libwayland-client at interpreter shutdown (16 of them in this spike's logs,
# all `python3.14 … segfault … libwayland-client.so`), *after* the result is
# printed — so a green run reports exit 139 and, worse, the compositor was
# observed crashing in the same second as a batch of these. The bytes are
# flushed; everything after this point is library cleanup we do not need.
sys.stdout.flush()
sys.stderr.flush()
os._exit(0)
if __name__ == "__main__":
sys.exit(main())
@@ -1,219 +0,0 @@
<?xml version="1.0" encoding="UTF-8"?>
<protocol name="ext_foreign_toplevel_list_v1">
<copyright>
Copyright © 2018 Ilia Bozhinov
Copyright © 2020 Isaac Freund
Copyright © 2022 wb9688
Copyright © 2023 i509VCB
Permission to use, copy, modify, distribute, and sell this
software and its documentation for any purpose is hereby granted
without fee, provided that the above copyright notice appear in
all copies and that both that copyright notice and this permission
notice appear in supporting documentation, and that the name of
the copyright holders not be used in advertising or publicity
pertaining to distribution of the software without specific,
written prior permission. The copyright holders make no
representations about the suitability of this software for any
purpose. It is provided "as is" without express or implied
warranty.
THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS
SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY
SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
THIS SOFTWARE.
</copyright>
<description summary="list toplevels">
The purpose of this protocol is to provide protocol object handles for
toplevels, possibly originating from another client.
This protocol is intentionally minimalistic and expects additional
functionality (e.g. creating a screencopy source from a toplevel handle,
getting information about the state of the toplevel) to be implemented
in extension protocols.
The compositor may choose to restrict this protocol to a special client
launched by the compositor itself or expose it to all clients,
this is compositor policy.
The key words "must", "must not", "required", "shall", "shall not",
"should", "should not", "recommended", "may", and "optional" in this
document are to be interpreted as described in IETF RFC 2119.
Warning! The protocol described in this file is currently in the testing
phase. Backward compatible changes may be added together with the
corresponding interface version bump. Backward incompatible changes can
only be done by creating a new major version of the extension.
</description>
<interface name="ext_foreign_toplevel_list_v1" version="1">
<description summary="list toplevels">
A toplevel is defined as a surface with a role similar to xdg_toplevel.
XWayland surfaces may be treated like toplevels in this protocol.
After a client binds the ext_foreign_toplevel_list_v1, each mapped
toplevel window will be sent using the ext_foreign_toplevel_list_v1.toplevel
event.
Clients which only care about the current state can perform a roundtrip after
binding this global.
For each instance of ext_foreign_toplevel_list_v1, the compositor must
create a new ext_foreign_toplevel_handle_v1 object for each mapped toplevel.
If a compositor implementation sends the ext_foreign_toplevel_list_v1.finished
event after the global is bound, the compositor must not send any
ext_foreign_toplevel_list_v1.toplevel events.
</description>
<event name="toplevel">
<description summary="a toplevel has been created">
This event is emitted whenever a new toplevel window is created. It is
emitted for all toplevels, regardless of the app that has created them.
All initial properties of the toplevel (identifier, title, app_id) will be sent
immediately after this event using the corresponding events for
ext_foreign_toplevel_handle_v1. The compositor will use the
ext_foreign_toplevel_handle_v1.done event to indicate when all data has
been sent.
</description>
<arg name="toplevel" type="new_id" interface="ext_foreign_toplevel_handle_v1"/>
</event>
<event name="finished">
<description summary="the compositor has finished with the toplevel manager">
This event indicates that the compositor is done sending events
to this object. The client should destroy the object.
See ext_foreign_toplevel_list_v1.destroy for more information.
The compositor must not send any more toplevel events after this event.
</description>
</event>
<request name="stop">
<description summary="stop sending events">
This request indicates that the client no longer wishes to receive
events for new toplevels.
The Wayland protocol is asynchronous, meaning the compositor may send
further toplevel events until the stop request is processed.
The client should wait for a ext_foreign_toplevel_list_v1.finished
event before destroying this object.
</description>
</request>
<request name="destroy" type="destructor">
<description summary="destroy the ext_foreign_toplevel_list_v1 object">
This request should be called either when the client will no longer
use the ext_foreign_toplevel_list_v1 or after the finished event
has been received to allow destruction of the object.
If a client wishes to destroy this object it should send a
ext_foreign_toplevel_list_v1.stop request and wait for a ext_foreign_toplevel_list_v1.finished
event, then destroy the handles and then this object.
</description>
</request>
</interface>
<interface name="ext_foreign_toplevel_handle_v1" version="1">
<description summary="a mapped toplevel">
A ext_foreign_toplevel_handle_v1 object represents a mapped toplevel
window. A single app may have multiple mapped toplevels.
</description>
<request name="destroy" type="destructor">
<description summary="destroy the ext_foreign_toplevel_handle_v1 object">
This request should be used when the client will no longer use the handle
or after the closed event has been received to allow destruction of the
object.
When a handle is destroyed, a new handle may not be created by the server
until the toplevel is unmapped and then remapped. Destroying a toplevel handle
is not recommended unless the client is cleaning up child objects
before destroying the ext_foreign_toplevel_list_v1 object, the toplevel
was closed or the toplevel handle will not be used in the future.
Other protocols which extend the ext_foreign_toplevel_handle_v1
interface should require destructors for extension interfaces be
called before allowing the toplevel handle to be destroyed.
</description>
</request>
<event name="closed">
<description summary="the toplevel has been closed">
The server will emit no further events on the ext_foreign_toplevel_handle_v1
after this event. Any requests received aside from the destroy request must
be ignored. Upon receiving this event, the client should destroy the handle.
Other protocols which extend the ext_foreign_toplevel_handle_v1
interface must also ignore requests other than destructors.
</description>
</event>
<event name="done">
<description summary="all information about the toplevel has been sent">
This event is sent after all changes in the toplevel state have
been sent.
This allows changes to the ext_foreign_toplevel_handle_v1 properties
to be atomically applied. Other protocols which extend the
ext_foreign_toplevel_handle_v1 interface may use this event to also
atomically apply any pending state.
This event must not be sent after the ext_foreign_toplevel_handle_v1.closed
event.
</description>
</event>
<event name="title">
<description summary="title change">
The title of the toplevel has changed.
The configured state must not be applied immediately. See
ext_foreign_toplevel_handle_v1.done for details.
</description>
<arg name="title" type="string"/>
</event>
<event name="app_id">
<description summary="app_id change">
The app id of the toplevel has changed.
The configured state must not be applied immediately. See
ext_foreign_toplevel_handle_v1.done for details.
</description>
<arg name="app_id" type="string"/>
</event>
<event name="identifier">
<description summary="a stable identifier for a toplevel">
This identifier is used to check if two or more toplevel handles belong
to the same toplevel.
The identifier is useful for command line tools or privileged clients
which may need to reference an exact toplevel across processes or
instances of the ext_foreign_toplevel_list_v1 global.
The compositor must only send this event when the handle is created.
The identifier must be unique per toplevel and its handles. Two different
toplevels must not have the same identifier. The identifier is only valid
as long as the toplevel is mapped. If the toplevel is unmapped the identifier
must not be reused. An identifier must not be reused by the compositor to
ensure there are no races when sharing identifiers between processes.
An identifier is a string that contains up to 32 printable ASCII bytes.
An identifier must not be an empty string. It is recommended that a
compositor includes an opaque generation value in identifiers. How the
generation value is used when generating the identifier is implementation
dependent.
</description>
<arg name="identifier" type="string"/>
</event>
</interface>
</protocol>
@@ -1,109 +0,0 @@
<?xml version="1.0" encoding="UTF-8"?>
<protocol name="ext_image_capture_source_v1">
<copyright>
Copyright © 2022 Andri Yngvason
Copyright © 2024 Simon Ser
Permission is hereby granted, free of charge, to any person obtaining a
copy of this software and associated documentation files (the "Software"),
to deal in the Software without restriction, including without limitation
the rights to use, copy, modify, merge, publish, distribute, sublicense,
and/or sell copies of the Software, and to permit persons to whom the
Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice (including the next
paragraph) shall be included in all copies or substantial portions of the
Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
DEALINGS IN THE SOFTWARE.
</copyright>
<description summary="opaque image capture source objects">
This protocol serves as an intermediary between capturing protocols and
potential image capture sources such as outputs and toplevels.
This protocol may be extended to support more image capture sources in the
future, thereby adding those image capture sources to other protocols that
use the image capture source object without having to modify those
protocols.
Warning! The protocol described in this file is currently in the testing
phase. Backward compatible changes may be added together with the
corresponding interface version bump. Backward incompatible changes can
only be done by creating a new major version of the extension.
</description>
<interface name="ext_image_capture_source_v1" version="1" frozen="true">
<description summary="opaque image capture source object">
The image capture source object is an opaque descriptor for a capturable
resource. This resource may be any sort of entity from which an image
may be derived.
Note, because ext_image_capture_source_v1 objects are created from multiple
independent factory interfaces, the ext_image_capture_source_v1 interface is
frozen at version 1.
</description>
<request name="destroy" type="destructor">
<description summary="delete this object">
Destroys the image capture source. This request may be sent at any time
by the client.
</description>
</request>
</interface>
<interface name="ext_output_image_capture_source_manager_v1" version="1">
<description summary="image capture source manager for outputs">
A manager for creating image capture source objects for wl_output objects.
</description>
<request name="create_source">
<description summary="create source object for output">
Creates a source object for an output. Images captured from this source
will show the same content as the output. Some elements may be omitted,
such as cursors and overlays that have been marked as transparent to
capturing.
</description>
<arg name="source" type="new_id" interface="ext_image_capture_source_v1"/>
<arg name="output" type="object" interface="wl_output"/>
</request>
<request name="destroy" type="destructor">
<description summary="delete this object">
Destroys the manager. This request may be sent at any time by the client
and objects created by the manager will remain valid after its
destruction.
</description>
</request>
</interface>
<interface name="ext_foreign_toplevel_image_capture_source_manager_v1" version="1">
<description summary="image capture source manager for foreign toplevels">
A manager for creating image capture source objects for
ext_foreign_toplevel_handle_v1 objects.
</description>
<request name="create_source">
<description summary="create source object for foreign toplevel">
Creates a source object for a foreign toplevel handle. Images captured
from this source will show the same content as the toplevel.
</description>
<arg name="source" type="new_id" interface="ext_image_capture_source_v1"/>
<arg name="toplevel_handle" type="object" interface="ext_foreign_toplevel_handle_v1"/>
</request>
<request name="destroy" type="destructor">
<description summary="delete this object">
Destroys the manager. This request may be sent at any time by the client
and objects created by the manager will remain valid after its
destruction.
</description>
</request>
</interface>
</protocol>
@@ -1,456 +0,0 @@
<?xml version="1.0" encoding="UTF-8"?>
<protocol name="ext_image_copy_capture_v1">
<copyright>
Copyright © 2021-2023 Andri Yngvason
Copyright © 2024 Simon Ser
Permission is hereby granted, free of charge, to any person obtaining a
copy of this software and associated documentation files (the "Software"),
to deal in the Software without restriction, including without limitation
the rights to use, copy, modify, merge, publish, distribute, sublicense,
and/or sell copies of the Software, and to permit persons to whom the
Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice (including the next
paragraph) shall be included in all copies or substantial portions of the
Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
DEALINGS IN THE SOFTWARE.
</copyright>
<description summary="image capturing into client buffers">
This protocol allows clients to ask the compositor to capture image sources
such as outputs and toplevels into user submitted buffers.
Warning! The protocol described in this file is currently in the testing
phase. Backward compatible changes may be added together with the
corresponding interface version bump. Backward incompatible changes can
only be done by creating a new major version of the extension.
</description>
<interface name="ext_image_copy_capture_manager_v1" version="1">
<description summary="manager to inform clients and begin capturing">
This object is a manager which offers requests to start capturing from a
source.
</description>
<enum name="error">
<entry name="invalid_option" value="1" summary="invalid option flag"/>
</enum>
<enum name="options" bitfield="true">
<entry name="paint_cursors" value="1" summary="paint cursors onto captured frames"/>
</enum>
<request name="create_session">
<description summary="capture an image capture source">
Create a capturing session for an image capture source.
If the paint_cursors option is set, cursors shall be composited onto
the captured frame. The cursor must not be composited onto the frame
if this flag is not set.
If the options bitfield is invalid, the invalid_option protocol error
is sent.
</description>
<arg name="session" type="new_id" interface="ext_image_copy_capture_session_v1"/>
<arg name="source" type="object" interface="ext_image_capture_source_v1"/>
<arg name="options" type="uint" enum="options"/>
</request>
<request name="create_pointer_cursor_session">
<description summary="capture the pointer cursor of an image capture source">
Create a cursor capturing session for the pointer of an image capture
source.
</description>
<arg name="session" type="new_id" interface="ext_image_copy_capture_cursor_session_v1"/>
<arg name="source" type="object" interface="ext_image_capture_source_v1"/>
<arg name="pointer" type="object" interface="wl_pointer"/>
</request>
<request name="destroy" type="destructor">
<description summary="destroy the manager">
Destroy the manager object.
Other objects created via this interface are unaffected.
</description>
</request>
</interface>
<interface name="ext_image_copy_capture_session_v1" version="1">
<description summary="image copy capture session">
This object represents an active image copy capture session.
After a capture session is created, buffer constraint events will be
emitted from the compositor to tell the client which buffer types and
formats are supported for reading from the session. The compositor may
re-send buffer constraint events whenever they change.
To advertise buffer constraints, the compositor must send in no
particular order: zero or more shm_format and dmabuf_format events, zero
or one dmabuf_device event, and exactly one buffer_size event. Then the
compositor must send a done event.
When the client has received all the buffer constraints, it can create a
buffer accordingly, attach it to the capture session using the
attach_buffer request, set the buffer damage using the damage_buffer
request and then send the capture request.
</description>
<enum name="error">
<entry name="duplicate_frame" value="1"
summary="create_frame sent before destroying previous frame"/>
</enum>
<event name="buffer_size">
<description summary="image capture source dimensions">
Provides the dimensions of the source image in buffer pixel coordinates.
The client must attach buffers that match this size.
</description>
<arg name="width" type="uint" summary="buffer width"/>
<arg name="height" type="uint" summary="buffer height"/>
</event>
<event name="shm_format">
<description summary="shm buffer format">
Provides the format that must be used for shared-memory buffers.
This event may be emitted multiple times, in which case the client may
choose any given format.
</description>
<arg name="format" type="uint" enum="wl_shm.format" summary="shm format"/>
</event>
<event name="dmabuf_device">
<description summary="dma-buf device">
This event advertises the device buffers must be allocated on for
dma-buf buffers.
In general the device is a DRM node. The DRM node type (primary vs.
render) is unspecified. Clients must not rely on the compositor sending
a particular node type. Clients cannot check two devices for equality
by comparing the dev_t value.
</description>
<arg name="device" type="array" summary="device dev_t value"/>
</event>
<event name="dmabuf_format">
<description summary="dma-buf format">
Provides the format that must be used for dma-buf buffers.
The client may choose any of the modifiers advertised in the array of
64-bit unsigned integers.
This event may be emitted multiple times, in which case the client may
choose any given format.
</description>
<arg name="format" type="uint" summary="drm format code"/>
<arg name="modifiers" type="array" summary="drm format modifiers"/>
</event>
<event name="done">
<description summary="all constraints have been sent">
This event is sent once when all buffer constraint events have been
sent.
The compositor must always end a batch of buffer constraint events with
this event, regardless of whether it sends the initial constraints or
an update.
</description>
</event>
<event name="stopped">
<description summary="session is no longer available">
This event indicates that the capture session has stopped and is no
longer available. This can happen in a number of cases, e.g. when the
underlying source is destroyed, if the user decides to end the image
capture, or if an unrecoverable runtime error has occurred.
The client should destroy the session after receiving this event.
</description>
</event>
<request name="create_frame">
<description summary="create a frame">
Create a capture frame for this session.
At most one frame object can exist for a given session at any time. If
a client sends a create_frame request before a previous frame object
has been destroyed, the duplicate_frame protocol error is raised.
</description>
<arg name="frame" type="new_id" interface="ext_image_copy_capture_frame_v1"/>
</request>
<request name="destroy" type="destructor">
<description summary="delete this object">
Destroys the session. This request can be sent at any time by the
client.
This request doesn't affect ext_image_copy_capture_frame_v1 objects created by
this object.
</description>
</request>
</interface>
<interface name="ext_image_copy_capture_frame_v1" version="1">
<description summary="image capture frame">
This object represents an image capture frame.
The client should attach a buffer, damage the buffer, and then send a
capture request.
If the capture is successful, the compositor must send the frame metadata
(transform, damage, presentation_time in any order) followed by the ready
event.
If the capture fails, the compositor must send the failed event.
</description>
<enum name="error">
<entry name="no_buffer" value="1" summary="capture sent without attach_buffer"/>
<entry name="invalid_buffer_damage" value="2" summary="invalid buffer damage"/>
<entry name="already_captured" value="3" summary="capture request has been sent"/>
</enum>
<request name="destroy" type="destructor">
<description summary="destroy this object">
Destroys the frame. This request can be sent at any time by the
client.
</description>
</request>
<request name="attach_buffer">
<description summary="attach buffer to session">
Attach a buffer to the session.
The wl_buffer.release request is unused.
The new buffer replaces any previously attached buffer.
This request must not be sent after capture, or else the
already_captured protocol error is raised.
</description>
<arg name="buffer" type="object" interface="wl_buffer"/>
</request>
<request name="damage_buffer">
<description summary="damage buffer">
Apply damage to the buffer which is to be captured next. This request
may be sent multiple times to describe a region.
The client indicates the accumulated damage since this wl_buffer was
last captured. During capture, the compositor will update the buffer
with at least the union of the region passed by the client and the
region advertised by ext_image_copy_capture_frame_v1.damage.
When a wl_buffer is captured for the first time, or when the client
doesn't track damage, the client must damage the whole buffer.
This is for optimisation purposes. The compositor may use this
information to reduce copying.
These coordinates originate from the upper left corner of the buffer.
If x or y are strictly negative, or if width or height are negative or
zero, the invalid_buffer_damage protocol error is raised.
This request must not be sent after capture, or else the
already_captured protocol error is raised.
</description>
<arg name="x" type="int" summary="region x coordinate"/>
<arg name="y" type="int" summary="region y coordinate"/>
<arg name="width" type="int" summary="region width"/>
<arg name="height" type="int" summary="region height"/>
</request>
<request name="capture">
<description summary="capture a frame">
Capture a frame.
Unless this is the first successful captured frame performed in this
session, the compositor may wait an indefinite amount of time for the
source content to change before performing the copy.
This request may only be sent once, or else the already_captured
protocol error is raised. A buffer must be attached before this request
is sent, or else the no_buffer protocol error is raised.
</description>
</request>
<event name="transform">
<description summary="buffer transform">
This event is sent before the ready event and holds the transform that
the compositor has applied to the buffer contents.
</description>
<arg name="transform" type="uint" enum="wl_output.transform"/>
</event>
<event name="damage">
<description summary="buffer damaged region">
This event is sent before the ready event. It may be generated multiple
times to describe a region.
The first captured frame in a session will always carry full damage.
Subsequent frames' damaged regions describe which parts of the buffer
have changed since the last ready event.
These coordinates originate in the upper left corner of the buffer.
</description>
<arg name="x" type="int" summary="damage x coordinate"/>
<arg name="y" type="int" summary="damage y coordinate"/>
<arg name="width" type="int" summary="damage width"/>
<arg name="height" type="int" summary="damage height"/>
</event>
<event name="presentation_time">
<description summary="presentation time of the frame">
This event indicates the time at which the frame is presented to the
output in system monotonic time. This event is sent before the ready
event.
The timestamp is expressed as tv_sec_hi, tv_sec_lo, tv_nsec triples,
each component being an unsigned 32-bit value. Whole seconds are in
tv_sec which is a 64-bit value combined from tv_sec_hi and tv_sec_lo,
and the additional fractional part in tv_nsec as nanoseconds. Hence,
for valid timestamps tv_nsec must be in [0, 999999999].
</description>
<arg name="tv_sec_hi" type="uint"
summary="high 32 bits of the seconds part of the timestamp"/>
<arg name="tv_sec_lo" type="uint"
summary="low 32 bits of the seconds part of the timestamp"/>
<arg name="tv_nsec" type="uint"
summary="nanoseconds part of the timestamp"/>
</event>
<event name="ready">
<description summary="frame is available for reading">
Called as soon as the frame is copied, indicating it is available
for reading.
The buffer may be re-used by the client after this event.
After receiving this event, the client must destroy the object.
</description>
</event>
<enum name="failure_reason">
<entry name="unknown" value="0">
<description summary="unknown runtime error">
An unspecified runtime error has occurred. The client may retry.
</description>
</entry>
<entry name="buffer_constraints" value="1">
<description summary="buffer constraints mismatch">
The buffer submitted by the client doesn't match the latest session
constraints. The client should re-allocate its buffers and retry.
</description>
</entry>
<entry name="stopped" value="2">
<description summary="session is no longer available">
The session has stopped. See ext_image_copy_capture_session_v1.stopped.
</description>
</entry>
</enum>
<event name="failed">
<description summary="capture failed">
This event indicates that the attempted frame copy has failed.
After receiving this event, the client must destroy the object.
</description>
<arg name="reason" type="uint" enum="failure_reason"/>
</event>
</interface>
<interface name="ext_image_copy_capture_cursor_session_v1" version="1">
<description summary="cursor capture session">
This object represents a cursor capture session. It extends the base
capture session with cursor-specific metadata.
</description>
<enum name="error">
<entry name="duplicate_session" value="1" summary="get_capture_session sent twice"/>
</enum>
<request name="destroy" type="destructor">
<description summary="delete this object">
Destroys the session. This request can be sent at any time by the
client.
This request doesn't affect ext_image_copy_capture_frame_v1 objects created by
this object.
</description>
</request>
<request name="get_capture_session">
<description summary="get image copy capturer session">
Gets the image copy capture session for this cursor session.
The session will produce frames of the cursor image. The compositor may
pause the session when the cursor leaves the captured area.
This request must not be sent more than once, or else the
duplicate_session protocol error is raised.
</description>
<arg name="session" type="new_id" interface="ext_image_copy_capture_session_v1"/>
</request>
<event name="enter">
<description summary="cursor entered captured area">
Sent when a cursor enters the captured area. It shall be generated
before the "position" and "hotspot" events when and only when a cursor
enters the area.
The cursor enters the captured area when the cursor image intersects
with the captured area. Note, this is different from e.g.
wl_pointer.enter.
</description>
</event>
<event name="leave">
<description summary="cursor left captured area">
Sent when a cursor leaves the captured area. No "position" or "hotspot"
event is generated for the cursor until the cursor enters the captured
area again.
</description>
</event>
<event name="position">
<description summary="position changed">
Cursors outside the image capture source do not get captured and no
event will be generated for them.
The given position is the position of the cursor's hotspot and it is
relative to the main buffer's top left corner in transformed buffer
pixel coordinates. The coordinates may be negative or greater than the
main buffer size.
</description>
<arg name="x" type="int" summary="position x coordinates"/>
<arg name="y" type="int" summary="position y coordinates"/>
</event>
<event name="hotspot">
<description summary="hotspot changed">
The hotspot describes the offset between the cursor image and the
position of the input device.
The given coordinates are the hotspot's offset from the origin in
buffer coordinates.
Clients should not apply the hotspot immediately: the hotspot becomes
effective when the next ext_image_copy_capture_frame_v1.ready event is received.
Compositors may delay this event until the client captures a new frame.
</description>
<arg name="x" type="int" summary="hotspot x coordinates"/>
<arg name="y" type="int" summary="hotspot y coordinates"/>
</event>
</interface>
</protocol>
@@ -1,232 +0,0 @@
<?xml version="1.0" encoding="UTF-8"?>
<protocol name="wlr_screencopy_unstable_v1">
<copyright>
Copyright © 2018 Simon Ser
Copyright © 2019 Andri Yngvason
Permission is hereby granted, free of charge, to any person obtaining a
copy of this software and associated documentation files (the "Software"),
to deal in the Software without restriction, including without limitation
the rights to use, copy, modify, merge, publish, distribute, sublicense,
and/or sell copies of the Software, and to permit persons to whom the
Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice (including the next
paragraph) shall be included in all copies or substantial portions of the
Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
DEALINGS IN THE SOFTWARE.
</copyright>
<description summary="screen content capturing on client buffers">
This protocol allows clients to ask the compositor to copy part of the
screen content to a client buffer.
Warning! The protocol described in this file is experimental and
backward incompatible changes may be made. Backward compatible changes
may be added together with the corresponding interface version bump.
Backward incompatible changes are done by bumping the version number in
the protocol and interface names and resetting the interface version.
Once the protocol is to be declared stable, the 'z' prefix and the
version number in the protocol and interface names are removed and the
interface version number is reset.
</description>
<interface name="zwlr_screencopy_manager_v1" version="3">
<description summary="manager to inform clients and begin capturing">
This object is a manager which offers requests to start capturing from a
source.
</description>
<request name="capture_output">
<description summary="capture an output">
Capture the next frame of an entire output.
</description>
<arg name="frame" type="new_id" interface="zwlr_screencopy_frame_v1"/>
<arg name="overlay_cursor" type="int"
summary="composite cursor onto the frame"/>
<arg name="output" type="object" interface="wl_output"/>
</request>
<request name="capture_output_region">
<description summary="capture an output's region">
Capture the next frame of an output's region.
The region is given in output logical coordinates, see
xdg_output.logical_size. The region will be clipped to the output's
extents.
</description>
<arg name="frame" type="new_id" interface="zwlr_screencopy_frame_v1"/>
<arg name="overlay_cursor" type="int"
summary="composite cursor onto the frame"/>
<arg name="output" type="object" interface="wl_output"/>
<arg name="x" type="int"/>
<arg name="y" type="int"/>
<arg name="width" type="int"/>
<arg name="height" type="int"/>
</request>
<request name="destroy" type="destructor">
<description summary="destroy the manager">
All objects created by the manager will still remain valid, until their
appropriate destroy request has been called.
</description>
</request>
</interface>
<interface name="zwlr_screencopy_frame_v1" version="3">
<description summary="a frame ready for copy">
This object represents a single frame.
When created, a series of buffer events will be sent, each representing a
supported buffer type. The "buffer_done" event is sent afterwards to
indicate that all supported buffer types have been enumerated. The client
will then be able to send a "copy" request. If the capture is successful,
the compositor will send a "flags" followed by a "ready" event.
For objects version 2 or lower, wl_shm buffers are always supported, ie.
the "buffer" event is guaranteed to be sent.
If the capture failed, the "failed" event is sent. This can happen anytime
before the "ready" event.
Once either a "ready" or a "failed" event is received, the client should
destroy the frame.
</description>
<event name="buffer">
<description summary="wl_shm buffer information">
Provides information about wl_shm buffer parameters that need to be
used for this frame. This event is sent once after the frame is created
if wl_shm buffers are supported.
</description>
<arg name="format" type="uint" enum="wl_shm.format" summary="buffer format"/>
<arg name="width" type="uint" summary="buffer width"/>
<arg name="height" type="uint" summary="buffer height"/>
<arg name="stride" type="uint" summary="buffer stride"/>
</event>
<request name="copy">
<description summary="copy the frame">
Copy the frame to the supplied buffer. The buffer must have a the
correct size, see zwlr_screencopy_frame_v1.buffer and
zwlr_screencopy_frame_v1.linux_dmabuf. The buffer needs to have a
supported format.
If the frame is successfully copied, a "flags" and a "ready" events are
sent. Otherwise, a "failed" event is sent.
</description>
<arg name="buffer" type="object" interface="wl_buffer"/>
</request>
<enum name="error">
<entry name="already_used" value="0"
summary="the object has already been used to copy a wl_buffer"/>
<entry name="invalid_buffer" value="1"
summary="buffer attributes are invalid"/>
</enum>
<enum name="flags" bitfield="true">
<entry name="y_invert" value="1" summary="contents are y-inverted"/>
</enum>
<event name="flags">
<description summary="frame flags">
Provides flags about the frame. This event is sent once before the
"ready" event.
</description>
<arg name="flags" type="uint" enum="flags" summary="frame flags"/>
</event>
<event name="ready">
<description summary="indicates frame is available for reading">
Called as soon as the frame is copied, indicating it is available
for reading. This event includes the time at which presentation happened
at.
The timestamp is expressed as tv_sec_hi, tv_sec_lo, tv_nsec triples,
each component being an unsigned 32-bit value. Whole seconds are in
tv_sec which is a 64-bit value combined from tv_sec_hi and tv_sec_lo,
and the additional fractional part in tv_nsec as nanoseconds. Hence,
for valid timestamps tv_nsec must be in [0, 999999999]. The seconds part
may have an arbitrary offset at start.
After receiving this event, the client should destroy the object.
</description>
<arg name="tv_sec_hi" type="uint"
summary="high 32 bits of the seconds part of the timestamp"/>
<arg name="tv_sec_lo" type="uint"
summary="low 32 bits of the seconds part of the timestamp"/>
<arg name="tv_nsec" type="uint"
summary="nanoseconds part of the timestamp"/>
</event>
<event name="failed">
<description summary="frame copy failed">
This event indicates that the attempted frame copy has failed.
After receiving this event, the client should destroy the object.
</description>
</event>
<request name="destroy" type="destructor">
<description summary="delete this object, used or not">
Destroys the frame. This request can be sent at any time by the client.
</description>
</request>
<!-- Version 2 additions -->
<request name="copy_with_damage" since="2">
<description summary="copy the frame when it's damaged">
Same as copy, except it waits until there is damage to copy.
</description>
<arg name="buffer" type="object" interface="wl_buffer"/>
</request>
<event name="damage" since="2">
<description summary="carries the coordinates of the damaged region">
This event is sent right before the ready event when copy_with_damage is
requested. It may be generated multiple times for each copy_with_damage
request.
The arguments describe a box around an area that has changed since the
last copy request that was derived from the current screencopy manager
instance.
The union of all regions received between the call to copy_with_damage
and a ready event is the total damage since the prior ready event.
</description>
<arg name="x" type="uint" summary="damaged x coordinates"/>
<arg name="y" type="uint" summary="damaged y coordinates"/>
<arg name="width" type="uint" summary="current width"/>
<arg name="height" type="uint" summary="current height"/>
</event>
<!-- Version 3 additions -->
<event name="linux_dmabuf" since="3">
<description summary="linux-dmabuf buffer information">
Provides information about linux-dmabuf buffer parameters that need to
be used for this frame. This event is sent once after the frame is
created if linux-dmabuf buffers are supported.
</description>
<arg name="format" type="uint" summary="fourcc pixel format"/>
<arg name="width" type="uint" summary="buffer width"/>
<arg name="height" type="uint" summary="buffer height"/>
</event>
<event name="buffer_done" since="3">
<description summary="all buffer types reported">
This event is sent once after all buffer events have been sent.
The client should proceed to create a buffer of one of the supported
types, and send a "copy" request.
</description>
</event>
</interface>
</protocol>
@@ -1,7 +0,0 @@
{
"window_ms": 10000.0,
"system_cpu_ms": 19910.0,
"compositor_cpu_ms": 9410.0,
"system_cpu_pct": 199.1,
"compositor_cpu_pct": 94.1
}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol ext-image-copy-capture-v1 (auto) mode jpeg target hh-chromium-gqkxz3 warmup 10 frames 200
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 107323 nonzero 105292
grimrs-capture: warmup 2/10 1280x720 bytes 107216 nonzero 105204
grimrs-capture: warmup 3/10 1280x720 bytes 107519 nonzero 105455
grimrs-capture: warmup 4/10 1280x720 bytes 107392 nonzero 105283
grimrs-capture: warmup 5/10 1280x720 bytes 107375 nonzero 105312
grimrs-capture: warmup 6/10 1280x720 bytes 107218 nonzero 105189
grimrs-capture: warmup 7/10 1280x720 bytes 107174 nonzero 105173
grimrs-capture: warmup 8/10 1280x720 bytes 107452 nonzero 105421
grimrs-capture: warmup 9/10 1280x720 bytes 107291 nonzero 105238
grimrs-capture: warmup 10/10 1280x720 bytes 107452 nonzero 105382
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"ext-image-copy-capture-v1 (auto)","output":"hh-chromium-gqkxz3","geometry":[1280,720],"mode":"jpeg","warmup":10,"frames":200,"frame_ms":{"min":61.074,"p50":124.039,"mean":132.54,"p95":198.748,"max":254.625},"fps_sustained":7.545,"wall_ms":26508.219,"cpu_ms_per_frame":6.6,"bytes_p50":107205.5,"persistent":true,"rss_peak_mb":746.25,"rss_peak_kb_raw":764156,"system_cpu_ms_per_frame":264.3,"compositor_cpu_ms_per_frame":123.1,"machine_cpu_pct":199.4,"frames_bytes_allocated_p50":3686400,"notes":"grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 105382."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol ext-image-copy-capture-v1 (auto) mode jpeg target hh-chromium-gqkxz3 warmup 10 frames 200
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 107394 nonzero 105367
grimrs-capture: warmup 2/10 1280x720 bytes 107137 nonzero 105116
grimrs-capture: warmup 3/10 1280x720 bytes 107298 nonzero 105216
grimrs-capture: warmup 4/10 1280x720 bytes 107092 nonzero 105064
grimrs-capture: warmup 5/10 1280x720 bytes 107031 nonzero 105045
grimrs-capture: warmup 6/10 1280x720 bytes 107334 nonzero 105280
grimrs-capture: warmup 7/10 1280x720 bytes 107182 nonzero 105134
grimrs-capture: warmup 8/10 1280x720 bytes 107395 nonzero 105345
grimrs-capture: warmup 9/10 1280x720 bytes 107333 nonzero 105326
grimrs-capture: warmup 10/10 1280x720 bytes 107105 nonzero 105080
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"ext-image-copy-capture-v1 (auto)","output":"hh-chromium-gqkxz3","geometry":[1280,720],"mode":"jpeg","warmup":10,"frames":200,"frame_ms":{"min":59.907,"p50":125.214,"mean":135.259,"p95":207.095,"max":324.014},"fps_sustained":7.393,"wall_ms":27051.989,"cpu_ms_per_frame":6.595,"bytes_p50":107203.5,"persistent":true,"rss_peak_mb":746.34,"rss_peak_kb_raw":764256,"system_cpu_ms_per_frame":269.1,"compositor_cpu_ms_per_frame":123.05,"machine_cpu_pct":199.0,"frames_bytes_allocated_p50":3686400,"notes":"grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 105080."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol ext-image-copy-capture-v1 (auto) mode raw target hh-chromium-gqkxz3 warmup 10 frames 200
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 2/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 3/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 4/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 5/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 6/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 7/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 8/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 9/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 10/10 1280x720 bytes 3686400 nonzero 3679565
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"ext-image-copy-capture-v1 (auto)","output":"hh-chromium-gqkxz3","geometry":[1280,720],"mode":"raw","warmup":10,"frames":200,"frame_ms":{"min":44.16,"p50":122.21,"mean":130.086,"p95":189.49,"max":226.412},"fps_sustained":7.687,"wall_ms":26017.423,"cpu_ms_per_frame":1.628,"bytes_p50":null,"persistent":true,"rss_peak_mb":746.09,"rss_peak_kb_raw":763992,"system_cpu_ms_per_frame":258.6,"compositor_cpu_ms_per_frame":120.85,"machine_cpu_pct":198.8,"frames_bytes_allocated_p50":3686400,"notes":"grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 3679565."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol ext-image-copy-capture-v1 (auto) mode raw target hh-chromium-gqkxz3 warmup 10 frames 200
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 2/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 3/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 4/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 5/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 6/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 7/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 8/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 9/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 10/10 1280x720 bytes 3686400 nonzero 3679565
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"ext-image-copy-capture-v1 (auto)","output":"hh-chromium-gqkxz3","geometry":[1280,720],"mode":"raw","warmup":10,"frames":200,"frame_ms":{"min":46.164,"p50":117.913,"mean":125.337,"p95":194.097,"max":255.125},"fps_sustained":7.978,"wall_ms":25067.547,"cpu_ms_per_frame":1.551,"bytes_p50":null,"persistent":true,"rss_peak_mb":746.02,"rss_peak_kb_raw":763924,"system_cpu_ms_per_frame":249.6,"compositor_cpu_ms_per_frame":117.6,"machine_cpu_pct":199.1,"frames_bytes_allocated_p50":3686400,"notes":"grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 3679565."}
Whitespace-only changes.
@@ -1 +0,0 @@
{"route": "grim-subprocess", "language": "python", "protocol": "zwlr-screencopy-v1 (grim's client)", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "jpeg", "warmup": 10, "frames": 200, "frame_ms": {"min": 58.928, "p50": 116.179, "mean": 121.352, "p95": 164.512, "max": 200.886}, "fps_sustained": 8.24, "wall_ms": 24614.05, "bytes_p50": 130597, "persistent": false, "notes": "one grim process per frame, fork and encode inside the timed span; territory: --output; jpeg q80", "cpu_ms_per_frame": 5.005, "rss_peak_mb": 18.47, "rss_peak_kb_raw": 18916, "system_cpu_ms_per_frame": 245.05, "compositor_cpu_ms_per_frame": 115.05, "machine_cpu_pct": 199.1, "machine_cpu_pct_of_all_cores": 99.6, "frames_bytes_allocated_p50": null, "resource_notes": ""}
Whitespace-only changes.
@@ -1 +0,0 @@
{"route": "grim-subprocess", "language": "python", "protocol": "zwlr-screencopy-v1 (grim's client)", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "jpeg", "warmup": 10, "frames": 200, "frame_ms": {"min": 67.203, "p50": 118.52, "mean": 124.224, "p95": 185.591, "max": 299.044}, "fps_sustained": 8.05, "wall_ms": 25182.14, "bytes_p50": 130590, "persistent": false, "notes": "one grim process per frame, fork and encode inside the timed span; territory: --output; jpeg q80", "cpu_ms_per_frame": 5.05, "rss_peak_mb": 18.45, "rss_peak_kb_raw": 18896, "system_cpu_ms_per_frame": 250.75, "compositor_cpu_ms_per_frame": 117.75, "machine_cpu_pct": 199.1, "machine_cpu_pct_of_all_cores": 99.6, "frames_bytes_allocated_p50": null, "resource_notes": ""}
Whitespace-only changes.
@@ -1 +0,0 @@
{"route": "grim-subprocess", "language": "python", "protocol": "zwlr-screencopy-v1 (grim's client)", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "raw", "warmup": 10, "frames": 200, "frame_ms": {"min": 51.756, "p50": 120.321, "mean": 125.705, "p95": 190.299, "max": 253.325}, "fps_sustained": 7.96, "wall_ms": 25452.51, "bytes_p50": 2764816, "persistent": false, "notes": "one grim process per frame, fork and encode inside the timed span; territory: --output; ppm, unencoded", "cpu_ms_per_frame": 3.856, "rss_peak_mb": 20.56, "rss_peak_kb_raw": 21052, "system_cpu_ms_per_frame": 253.8, "compositor_cpu_ms_per_frame": 118.55, "machine_cpu_pct": 199.4, "machine_cpu_pct_of_all_cores": 99.7, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
Whitespace-only changes.
@@ -1 +0,0 @@
{"route": "grim-subprocess", "language": "python", "protocol": "zwlr-screencopy-v1 (grim's client)", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "raw", "warmup": 10, "frames": 200, "frame_ms": {"min": 64.061, "p50": 122.389, "mean": 127.307, "p95": 196.38, "max": 264.387}, "fps_sustained": 7.86, "wall_ms": 25757.6, "bytes_p50": 2764816, "persistent": false, "notes": "one grim process per frame, fork and encode inside the timed span; territory: --output; ppm, unencoded", "cpu_ms_per_frame": 3.858, "rss_peak_mb": 20.51, "rss_peak_kb_raw": 21004, "system_cpu_ms_per_frame": 256.65, "compositor_cpu_ms_per_frame": 118.25, "machine_cpu_pct": 199.3, "machine_cpu_pct_of_all_cores": 99.6, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
Whitespace-only changes.
@@ -1 +0,0 @@
{"route": "pywayland-in-process", "language": "python", "protocol": "ext-image-copy-capture-v1", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "jpeg", "warmup": 10, "frames": 200, "frame_ms": {"min": 30.44, "p50": 76.633, "mean": 71.153, "p95": 103.055, "max": 185.942}, "stage_ms": {"capture": 68.35, "encode": 2.708, "stitch": 0.0}, "fps_sustained": 14.05, "wall_ms": 14231.13, "bytes_p50": 105794, "persistent": true, "notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame", "cpu_ms_per_frame": 1.811, "rss_peak_mb": 28.99, "rss_peak_kb_raw": 29688, "system_cpu_ms_per_frame": 141.75, "compositor_cpu_ms_per_frame": 66.55, "machine_cpu_pct": 199.2, "machine_cpu_pct_of_all_cores": 99.6, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
Whitespace-only changes.
@@ -1 +0,0 @@
{"route": "pywayland-in-process", "language": "python", "protocol": "ext-image-copy-capture-v1", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "jpeg", "warmup": 10, "frames": 200, "frame_ms": {"min": 38.593, "p50": 78.693, "mean": 71.507, "p95": 96.637, "max": 129.106}, "stage_ms": {"capture": 69.115, "encode": 2.339, "stitch": 0.0}, "fps_sustained": 13.98, "wall_ms": 14301.79, "bytes_p50": 105804, "persistent": true, "notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame", "cpu_ms_per_frame": 1.776, "rss_peak_mb": 29.25, "rss_peak_kb_raw": 29956, "system_cpu_ms_per_frame": 142.65, "compositor_cpu_ms_per_frame": 67.9, "machine_cpu_pct": 199.5, "machine_cpu_pct_of_all_cores": 99.7, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
Whitespace-only changes.
@@ -1 +0,0 @@
{"route": "pywayland-in-process", "language": "python", "protocol": "ext-image-copy-capture-v1", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "raw", "warmup": 10, "frames": 200, "frame_ms": {"min": 22.685, "p50": 71.372, "mean": 69.708, "p95": 89.816, "max": 121.781}, "stage_ms": {"capture": 69.698, "encode": 0.0, "stitch": 0.0}, "fps_sustained": 14.35, "wall_ms": 13941.78, "bytes_p50": null, "persistent": true, "notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame", "cpu_ms_per_frame": 0.311, "rss_peak_mb": 24.43, "rss_peak_kb_raw": 25020, "system_cpu_ms_per_frame": 138.6, "compositor_cpu_ms_per_frame": 65.85, "machine_cpu_pct": 198.8, "machine_cpu_pct_of_all_cores": 99.4, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
Whitespace-only changes.
@@ -1 +0,0 @@
{"route": "pywayland-in-process", "language": "python", "protocol": "ext-image-copy-capture-v1", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "raw", "warmup": 10, "frames": 200, "frame_ms": {"min": 41.241, "p50": 75.703, "mean": 69.499, "p95": 86.611, "max": 112.812}, "stage_ms": {"capture": 69.489, "encode": 0.0, "stitch": 0.0}, "fps_sustained": 14.39, "wall_ms": 13899.94, "bytes_p50": null, "persistent": true, "notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame", "cpu_ms_per_frame": 0.318, "rss_peak_mb": 24.52, "rss_peak_kb_raw": 25112, "system_cpu_ms_per_frame": 138.35, "compositor_cpu_ms_per_frame": 66.45, "machine_cpu_pct": 199.1, "machine_cpu_pct_of_all_cores": 99.5, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
@@ -1 +0,0 @@
{"route": "pywayland-in-process", "language": "python", "protocol": "zwlr-screencopy-v1", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "jpeg", "warmup": 10, "frames": 200, "frame_ms": {"min": 25.597, "p50": 79.088, "mean": 71.771, "p95": 96.766, "max": 132.204}, "stage_ms": {"capture": 69.093, "encode": 2.522, "stitch": 0.0}, "fps_sustained": 13.93, "wall_ms": 14354.55, "bytes_p50": 105799, "persistent": true, "notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame", "cpu_ms_per_frame": 1.781, "rss_peak_mb": 29.38, "rss_peak_kb_raw": 30084, "system_cpu_ms_per_frame": 143.0, "compositor_cpu_ms_per_frame": 67.4, "machine_cpu_pct": 199.2, "machine_cpu_pct_of_all_cores": 99.6, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
@@ -1 +0,0 @@
{"route": "pywayland-in-process", "language": "python", "protocol": "zwlr-screencopy-v1", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "jpeg", "warmup": 10, "frames": 200, "frame_ms": {"min": 39.218, "p50": 76.971, "mean": 71.88, "p95": 95.795, "max": 149.48}, "stage_ms": {"capture": 69.53, "encode": 2.305, "stitch": 0.0}, "fps_sustained": 13.91, "wall_ms": 14378.78, "bytes_p50": 105813, "persistent": true, "notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame", "cpu_ms_per_frame": 1.782, "rss_peak_mb": 29.3, "rss_peak_kb_raw": 30000, "system_cpu_ms_per_frame": 143.05, "compositor_cpu_ms_per_frame": 67.55, "machine_cpu_pct": 199.0, "machine_cpu_pct_of_all_cores": 99.5, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
@@ -1 +0,0 @@
{"route": "pywayland-in-process", "language": "python", "protocol": "zwlr-screencopy-v1", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "raw", "warmup": 10, "frames": 200, "frame_ms": {"min": 40.891, "p50": 76.175, "mean": 70.827, "p95": 96.846, "max": 125.471}, "stage_ms": {"capture": 70.816, "encode": 0.0, "stitch": 0.0}, "fps_sustained": 14.12, "wall_ms": 14165.47, "bytes_p50": null, "persistent": true, "notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame", "cpu_ms_per_frame": 0.33, "rss_peak_mb": 24.45, "rss_peak_kb_raw": 25036, "system_cpu_ms_per_frame": 140.65, "compositor_cpu_ms_per_frame": 67.3, "machine_cpu_pct": 198.6, "machine_cpu_pct_of_all_cores": 99.3, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
@@ -1 +0,0 @@
{"route": "pywayland-in-process", "language": "python", "protocol": "zwlr-screencopy-v1", "output": "hh-chromium-gqkxz3", "geometry": [1280, 720], "mode": "raw", "warmup": 10, "frames": 200, "frame_ms": {"min": 42.575, "p50": 77.219, "mean": 70.469, "p95": 93.935, "max": 122.928}, "stage_ms": {"capture": 70.459, "encode": 0.0, "stitch": 0.0}, "fps_sustained": 14.19, "wall_ms": 14094.03, "bytes_p50": null, "persistent": true, "notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame", "cpu_ms_per_frame": 0.328, "rss_peak_mb": 24.45, "rss_peak_kb_raw": 25036, "system_cpu_ms_per_frame": 140.1, "compositor_cpu_ms_per_frame": 66.95, "machine_cpu_pct": 198.8, "machine_cpu_pct_of_all_cores": 99.4, "frames_bytes_allocated_p50": 3686400, "resource_notes": ""}
@@ -1,674 +0,0 @@
[
{
"route": "grim-subprocess",
"description": "grim subprocess per frame (shipped)",
"cell": "output-raw",
"run": 1,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/grim_producer.py --output hh-chromium-gqkxz3 --frames 200 --warmup 10 --mode raw --json",
"rc": 0,
"elapsed_s": 26.8,
"language": "python",
"protocol": "zwlr-screencopy-v1 (grim's client)",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "raw",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 51.756,
"p50": 120.321,
"mean": 125.705,
"p95": 190.299,
"max": 253.325
},
"fps_sustained": 7.96,
"wall_ms": 25452.51,
"bytes_p50": 2764816,
"persistent": false,
"notes": "one grim process per frame, fork and encode inside the timed span; territory: --output; ppm, unencoded",
"cpu_ms_per_frame": 3.856,
"rss_peak_mb": 20.56,
"rss_peak_kb_raw": 21052,
"system_cpu_ms_per_frame": 253.8,
"compositor_cpu_ms_per_frame": 118.55,
"machine_cpu_pct": 199.4,
"machine_cpu_pct_of_all_cores": 99.7,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "grim-subprocess",
"description": "grim subprocess per frame (shipped)",
"cell": "output-raw",
"run": 2,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/grim_producer.py --output hh-chromium-gqkxz3 --frames 200 --warmup 10 --mode raw --json",
"rc": 0,
"elapsed_s": 27.0,
"language": "python",
"protocol": "zwlr-screencopy-v1 (grim's client)",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "raw",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 64.061,
"p50": 122.389,
"mean": 127.307,
"p95": 196.38,
"max": 264.387
},
"fps_sustained": 7.86,
"wall_ms": 25757.6,
"bytes_p50": 2764816,
"persistent": false,
"notes": "one grim process per frame, fork and encode inside the timed span; territory: --output; ppm, unencoded",
"cpu_ms_per_frame": 3.858,
"rss_peak_mb": 20.51,
"rss_peak_kb_raw": 21004,
"system_cpu_ms_per_frame": 256.65,
"compositor_cpu_ms_per_frame": 118.25,
"machine_cpu_pct": 199.3,
"machine_cpu_pct_of_all_cores": 99.6,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "grim-subprocess",
"description": "grim subprocess per frame (shipped)",
"cell": "output-jpeg",
"run": 1,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/grim_producer.py --output hh-chromium-gqkxz3 --frames 200 --warmup 10 --mode jpeg --json",
"rc": 0,
"elapsed_s": 26.1,
"language": "python",
"protocol": "zwlr-screencopy-v1 (grim's client)",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "jpeg",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 58.928,
"p50": 116.179,
"mean": 121.352,
"p95": 164.512,
"max": 200.886
},
"fps_sustained": 8.24,
"wall_ms": 24614.05,
"bytes_p50": 130597,
"persistent": false,
"notes": "one grim process per frame, fork and encode inside the timed span; territory: --output; jpeg q80",
"cpu_ms_per_frame": 5.005,
"rss_peak_mb": 18.47,
"rss_peak_kb_raw": 18916,
"system_cpu_ms_per_frame": 245.05,
"compositor_cpu_ms_per_frame": 115.05,
"machine_cpu_pct": 199.1,
"machine_cpu_pct_of_all_cores": 99.6,
"frames_bytes_allocated_p50": null,
"resource_notes": ""
},
{
"route": "grim-subprocess",
"description": "grim subprocess per frame (shipped)",
"cell": "output-jpeg",
"run": 2,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/grim_producer.py --output hh-chromium-gqkxz3 --frames 200 --warmup 10 --mode jpeg --json",
"rc": 0,
"elapsed_s": 26.6,
"language": "python",
"protocol": "zwlr-screencopy-v1 (grim's client)",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "jpeg",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 67.203,
"p50": 118.52,
"mean": 124.224,
"p95": 185.591,
"max": 299.044
},
"fps_sustained": 8.05,
"wall_ms": 25182.14,
"bytes_p50": 130590,
"persistent": false,
"notes": "one grim process per frame, fork and encode inside the timed span; territory: --output; jpeg q80",
"cpu_ms_per_frame": 5.05,
"rss_peak_mb": 18.45,
"rss_peak_kb_raw": 18896,
"system_cpu_ms_per_frame": 250.75,
"compositor_cpu_ms_per_frame": 117.75,
"machine_cpu_pct": 199.1,
"machine_cpu_pct_of_all_cores": 99.6,
"frames_bytes_allocated_p50": null,
"resource_notes": ""
},
{
"route": "pywayland-in-process",
"description": "pywayland in-process, zwlr-screencopy-v1",
"cell": "output-raw",
"run": 1,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/wayland_producer.py --output hh-chromium-gqkxz3 --protocol screencopy --frames 200 --warmup 10 --mode raw --json",
"rc": 0,
"elapsed_s": 15.2,
"language": "python",
"protocol": "zwlr-screencopy-v1",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "raw",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 40.891,
"p50": 76.175,
"mean": 70.827,
"p95": 96.846,
"max": 125.471
},
"stage_ms": {
"capture": 70.816,
"encode": 0.0,
"stitch": 0.0
},
"fps_sustained": 14.12,
"wall_ms": 14165.47,
"bytes_p50": null,
"persistent": true,
"notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame",
"cpu_ms_per_frame": 0.33,
"rss_peak_mb": 24.45,
"rss_peak_kb_raw": 25036,
"system_cpu_ms_per_frame": 140.65,
"compositor_cpu_ms_per_frame": 67.3,
"machine_cpu_pct": 198.6,
"machine_cpu_pct_of_all_cores": 99.3,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "pywayland-in-process",
"description": "pywayland in-process, zwlr-screencopy-v1",
"cell": "output-raw",
"run": 2,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/wayland_producer.py --output hh-chromium-gqkxz3 --protocol screencopy --frames 200 --warmup 10 --mode raw --json",
"rc": 0,
"elapsed_s": 15.1,
"language": "python",
"protocol": "zwlr-screencopy-v1",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "raw",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 42.575,
"p50": 77.219,
"mean": 70.469,
"p95": 93.935,
"max": 122.928
},
"stage_ms": {
"capture": 70.459,
"encode": 0.0,
"stitch": 0.0
},
"fps_sustained": 14.19,
"wall_ms": 14094.03,
"bytes_p50": null,
"persistent": true,
"notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame",
"cpu_ms_per_frame": 0.328,
"rss_peak_mb": 24.45,
"rss_peak_kb_raw": 25036,
"system_cpu_ms_per_frame": 140.1,
"compositor_cpu_ms_per_frame": 66.95,
"machine_cpu_pct": 198.8,
"machine_cpu_pct_of_all_cores": 99.4,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "pywayland-in-process",
"description": "pywayland in-process, zwlr-screencopy-v1",
"cell": "output-jpeg",
"run": 1,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/wayland_producer.py --output hh-chromium-gqkxz3 --protocol screencopy --frames 200 --warmup 10 --mode jpeg --json",
"rc": 0,
"elapsed_s": 15.4,
"language": "python",
"protocol": "zwlr-screencopy-v1",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "jpeg",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 25.597,
"p50": 79.088,
"mean": 71.771,
"p95": 96.766,
"max": 132.204
},
"stage_ms": {
"capture": 69.093,
"encode": 2.522,
"stitch": 0.0
},
"fps_sustained": 13.93,
"wall_ms": 14354.55,
"bytes_p50": 105799,
"persistent": true,
"notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame",
"cpu_ms_per_frame": 1.781,
"rss_peak_mb": 29.38,
"rss_peak_kb_raw": 30084,
"system_cpu_ms_per_frame": 143.0,
"compositor_cpu_ms_per_frame": 67.4,
"machine_cpu_pct": 199.2,
"machine_cpu_pct_of_all_cores": 99.6,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "pywayland-in-process",
"description": "pywayland in-process, zwlr-screencopy-v1",
"cell": "output-jpeg",
"run": 2,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/wayland_producer.py --output hh-chromium-gqkxz3 --protocol screencopy --frames 200 --warmup 10 --mode jpeg --json",
"rc": 0,
"elapsed_s": 15.4,
"language": "python",
"protocol": "zwlr-screencopy-v1",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "jpeg",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 39.218,
"p50": 76.971,
"mean": 71.88,
"p95": 95.795,
"max": 149.48
},
"stage_ms": {
"capture": 69.53,
"encode": 2.305,
"stitch": 0.0
},
"fps_sustained": 13.91,
"wall_ms": 14378.78,
"bytes_p50": 105813,
"persistent": true,
"notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame",
"cpu_ms_per_frame": 1.782,
"rss_peak_mb": 29.3,
"rss_peak_kb_raw": 30000,
"system_cpu_ms_per_frame": 143.05,
"compositor_cpu_ms_per_frame": 67.55,
"machine_cpu_pct": 199.0,
"machine_cpu_pct_of_all_cores": 99.5,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "pywayland-in-process",
"description": "pywayland in-process, ext-image-copy-capture-v1",
"cell": "output-raw",
"run": 1,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/wayland_producer.py --output hh-chromium-gqkxz3 --protocol ext --frames 200 --warmup 10 --mode raw --json",
"rc": 0,
"elapsed_s": 15.0,
"language": "python",
"protocol": "ext-image-copy-capture-v1",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "raw",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 22.685,
"p50": 71.372,
"mean": 69.708,
"p95": 89.816,
"max": 121.781
},
"stage_ms": {
"capture": 69.698,
"encode": 0.0,
"stitch": 0.0
},
"fps_sustained": 14.35,
"wall_ms": 13941.78,
"bytes_p50": null,
"persistent": true,
"notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame",
"cpu_ms_per_frame": 0.311,
"rss_peak_mb": 24.43,
"rss_peak_kb_raw": 25020,
"system_cpu_ms_per_frame": 138.6,
"compositor_cpu_ms_per_frame": 65.85,
"machine_cpu_pct": 198.8,
"machine_cpu_pct_of_all_cores": 99.4,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "pywayland-in-process",
"description": "pywayland in-process, ext-image-copy-capture-v1",
"cell": "output-raw",
"run": 2,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/wayland_producer.py --output hh-chromium-gqkxz3 --protocol ext --frames 200 --warmup 10 --mode raw --json",
"rc": 0,
"elapsed_s": 14.9,
"language": "python",
"protocol": "ext-image-copy-capture-v1",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "raw",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 41.241,
"p50": 75.703,
"mean": 69.499,
"p95": 86.611,
"max": 112.812
},
"stage_ms": {
"capture": 69.489,
"encode": 0.0,
"stitch": 0.0
},
"fps_sustained": 14.39,
"wall_ms": 13899.94,
"bytes_p50": null,
"persistent": true,
"notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame",
"cpu_ms_per_frame": 0.318,
"rss_peak_mb": 24.52,
"rss_peak_kb_raw": 25112,
"system_cpu_ms_per_frame": 138.35,
"compositor_cpu_ms_per_frame": 66.45,
"machine_cpu_pct": 199.1,
"machine_cpu_pct_of_all_cores": 99.5,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "pywayland-in-process",
"description": "pywayland in-process, ext-image-copy-capture-v1",
"cell": "output-jpeg",
"run": 1,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/wayland_producer.py --output hh-chromium-gqkxz3 --protocol ext --frames 200 --warmup 10 --mode jpeg --json",
"rc": 0,
"elapsed_s": 15.3,
"language": "python",
"protocol": "ext-image-copy-capture-v1",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "jpeg",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 30.44,
"p50": 76.633,
"mean": 71.153,
"p95": 103.055,
"max": 185.942
},
"stage_ms": {
"capture": 68.35,
"encode": 2.708,
"stitch": 0.0
},
"fps_sustained": 14.05,
"wall_ms": 14231.13,
"bytes_p50": 105794,
"persistent": true,
"notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame",
"cpu_ms_per_frame": 1.811,
"rss_peak_mb": 28.99,
"rss_peak_kb_raw": 29688,
"system_cpu_ms_per_frame": 141.75,
"compositor_cpu_ms_per_frame": 66.55,
"machine_cpu_pct": 199.2,
"machine_cpu_pct_of_all_cores": 99.6,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "pywayland-in-process",
"description": "pywayland in-process, ext-image-copy-capture-v1",
"cell": "output-jpeg",
"run": 2,
"command": "/nix/store/i7d46siafmlrgcmj9nzvwj7y64p34yaz-python3-3.14.7-env/bin/python3 /var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/producers/wayland_producer.py --output hh-chromium-gqkxz3 --protocol ext --frames 200 --warmup 10 --mode jpeg --json",
"rc": 0,
"elapsed_s": 15.3,
"language": "python",
"protocol": "ext-image-copy-capture-v1",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "jpeg",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 38.593,
"p50": 78.693,
"mean": 71.507,
"p95": 96.637,
"max": 129.106
},
"stage_ms": {
"capture": 69.115,
"encode": 2.339,
"stitch": 0.0
},
"fps_sustained": 13.98,
"wall_ms": 14301.79,
"bytes_p50": 105804,
"persistent": true,
"notes": "one connection, buffers reused, turbojpeg q80 in-process; 1 output(s); forcing a repaint per frame",
"cpu_ms_per_frame": 1.776,
"rss_peak_mb": 29.25,
"rss_peak_kb_raw": 29956,
"system_cpu_ms_per_frame": 142.65,
"compositor_cpu_ms_per_frame": 67.9,
"machine_cpu_pct": 199.5,
"machine_cpu_pct_of_all_cores": 99.7,
"frames_bytes_allocated_p50": 3686400,
"resource_notes": ""
},
{
"route": "grim-rs",
"description": "grim-rs crate (Rust), long-lived process",
"cell": "output-raw",
"run": 1,
"command": "/var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/001a-grim-rs/target/release/grimrs-capture --output hh-chromium-gqkxz3 --frames 200 --warmup 10 --mode raw --json",
"rc": 0,
"elapsed_s": 27.7,
"language": "rust",
"protocol": "ext-image-copy-capture-v1 (auto)",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "raw",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 44.16,
"p50": 122.21,
"mean": 130.086,
"p95": 189.49,
"max": 226.412
},
"fps_sustained": 7.687,
"wall_ms": 26017.423,
"cpu_ms_per_frame": 1.628,
"bytes_p50": null,
"persistent": true,
"rss_peak_mb": 746.09,
"rss_peak_kb_raw": 763992,
"system_cpu_ms_per_frame": 258.6,
"compositor_cpu_ms_per_frame": 120.85,
"machine_cpu_pct": 198.8,
"frames_bytes_allocated_p50": 3686400,
"notes": "grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 3679565."
},
{
"route": "grim-rs",
"description": "grim-rs crate (Rust), long-lived process",
"cell": "output-raw",
"run": 2,
"command": "/var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/001a-grim-rs/target/release/grimrs-capture --output hh-chromium-gqkxz3 --frames 200 --warmup 10 --mode raw --json",
"rc": 0,
"elapsed_s": 26.7,
"language": "rust",
"protocol": "ext-image-copy-capture-v1 (auto)",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "raw",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 46.164,
"p50": 117.913,
"mean": 125.337,
"p95": 194.097,
"max": 255.125
},
"fps_sustained": 7.978,
"wall_ms": 25067.547,
"cpu_ms_per_frame": 1.551,
"bytes_p50": null,
"persistent": true,
"rss_peak_mb": 746.02,
"rss_peak_kb_raw": 763924,
"system_cpu_ms_per_frame": 249.6,
"compositor_cpu_ms_per_frame": 117.6,
"machine_cpu_pct": 199.1,
"frames_bytes_allocated_p50": 3686400,
"notes": "grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 3679565."
},
{
"route": "grim-rs",
"description": "grim-rs crate (Rust), long-lived process",
"cell": "output-jpeg",
"run": 1,
"command": "/var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/001a-grim-rs/target/release/grimrs-capture --output hh-chromium-gqkxz3 --frames 200 --warmup 10 --mode jpeg --json",
"rc": 0,
"elapsed_s": 28.2,
"language": "rust",
"protocol": "ext-image-copy-capture-v1 (auto)",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "jpeg",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 61.074,
"p50": 124.039,
"mean": 132.54,
"p95": 198.748,
"max": 254.625
},
"fps_sustained": 7.545,
"wall_ms": 26508.219,
"cpu_ms_per_frame": 6.6,
"bytes_p50": 107205.5,
"persistent": true,
"rss_peak_mb": 746.25,
"rss_peak_kb_raw": 764156,
"system_cpu_ms_per_frame": 264.3,
"compositor_cpu_ms_per_frame": 123.1,
"machine_cpu_pct": 199.4,
"frames_bytes_allocated_p50": 3686400,
"notes": "grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 105382."
},
{
"route": "grim-rs",
"description": "grim-rs crate (Rust), long-lived process",
"cell": "output-jpeg",
"run": 2,
"command": "/var/lib/hermes/worktrees/hhs-capture-speed/spikes/capture-speed/001a-grim-rs/target/release/grimrs-capture --output hh-chromium-gqkxz3 --frames 200 --warmup 10 --mode jpeg --json",
"rc": 0,
"elapsed_s": 28.6,
"language": "rust",
"protocol": "ext-image-copy-capture-v1 (auto)",
"output": "hh-chromium-gqkxz3",
"geometry": [
1280,
720
],
"mode": "jpeg",
"warmup": 10,
"frames": 200,
"frame_ms": {
"min": 59.907,
"p50": 125.214,
"mean": 135.259,
"p95": 207.095,
"max": 324.014
},
"fps_sustained": 7.393,
"wall_ms": 27051.989,
"cpu_ms_per_frame": 6.595,
"bytes_p50": 107203.5,
"persistent": true,
"rss_peak_mb": 746.34,
"rss_peak_kb_raw": 764256,
"system_cpu_ms_per_frame": 269.1,
"compositor_cpu_ms_per_frame": 123.05,
"machine_cpu_pct": 199.0,
"frames_bytes_allocated_p50": 3686400,
"notes": "grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 105080."
}
]
@@ -1,7 +0,0 @@
{
"window_ms": 10000.0,
"system_cpu_ms": 19850.0,
"compositor_cpu_ms": 9500.0,
"system_cpu_pct": 198.5,
"compositor_cpu_pct": 95.0
}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol ext-image-copy-capture-v1 (auto) mode jpeg target hh-chromium-52w2dm warmup 10 frames 200
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1; hh-chromium-52w2dm@2560,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 2/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 3/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 4/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 5/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 6/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 7/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 8/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 9/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 10/10 1280x720 bytes 107829 nonzero 105852
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"ext-image-copy-capture-v1 (auto)","output":"hh-chromium-52w2dm","geometry":[1280,720],"mode":"jpeg","warmup":10,"frames":200,"frame_ms":{"min":95.334,"p50":186.693,"mean":199.129,"p95":282.41,"max":411.263},"fps_sustained":5.022,"wall_ms":39825.916,"cpu_ms_per_frame":6.7,"bytes_p50":107829.0,"persistent":true,"rss_peak_mb":746.3,"rss_peak_kb_raw":764208,"system_cpu_ms_per_frame":396.0,"compositor_cpu_ms_per_frame":187.75,"machine_cpu_pct":198.9,"frames_bytes_allocated_p50":3686400,"notes":"grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1; hh-chromium-52w2dm@2560,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 105852."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol ext-image-copy-capture-v1 (auto) mode jpeg target hh-chromium-52w2dm warmup 10 frames 200
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1; hh-chromium-52w2dm@2560,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 2/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 3/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 4/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 5/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 6/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 7/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 8/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 9/10 1280x720 bytes 107829 nonzero 105852
grimrs-capture: warmup 10/10 1280x720 bytes 107829 nonzero 105852
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"ext-image-copy-capture-v1 (auto)","output":"hh-chromium-52w2dm","geometry":[1280,720],"mode":"jpeg","warmup":10,"frames":200,"frame_ms":{"min":104.636,"p50":184.629,"mean":198.848,"p95":306.666,"max":368.522},"fps_sustained":5.029,"wall_ms":39771.382,"cpu_ms_per_frame":6.67,"bytes_p50":107829.0,"persistent":true,"rss_peak_mb":746.35,"rss_peak_kb_raw":764264,"system_cpu_ms_per_frame":395.55,"compositor_cpu_ms_per_frame":187.5,"machine_cpu_pct":198.9,"frames_bytes_allocated_p50":3686400,"notes":"grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1; hh-chromium-52w2dm@2560,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 105852."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol ext-image-copy-capture-v1 (auto) mode raw target hh-chromium-52w2dm warmup 10 frames 200
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1; hh-chromium-52w2dm@2560,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 2/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 3/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 4/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 5/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 6/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 7/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 8/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 9/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 10/10 1280x720 bytes 3686400 nonzero 3679565
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"ext-image-copy-capture-v1 (auto)","output":"hh-chromium-52w2dm","geometry":[1280,720],"mode":"raw","warmup":10,"frames":200,"frame_ms":{"min":112.37,"p50":188.581,"mean":199.48,"p95":289.422,"max":358.439},"fps_sustained":5.013,"wall_ms":39896.261,"cpu_ms_per_frame":1.677,"bytes_p50":null,"persistent":true,"rss_peak_mb":746.18,"rss_peak_kb_raw":764084,"system_cpu_ms_per_frame":396.45,"compositor_cpu_ms_per_frame":188.85,"machine_cpu_pct":198.7,"frames_bytes_allocated_p50":3686400,"notes":"grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1; hh-chromium-52w2dm@2560,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 3679565."}
@@ -1,12 +0,0 @@
grimrs-capture: route grim-rs protocol ext-image-copy-capture-v1 (auto) mode raw target hh-chromium-52w2dm warmup 10 frames 200
grimrs-capture: outputs Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1; hh-chromium-52w2dm@2560,0 1280x720 scale 1
grimrs-capture: warmup 1/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 2/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 3/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 4/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 5/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 6/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 7/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 8/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 9/10 1280x720 bytes 3686400 nonzero 3679565
grimrs-capture: warmup 10/10 1280x720 bytes 3686400 nonzero 3679565
@@ -1 +0,0 @@
{"route":"grim-rs","language":"rust","protocol":"ext-image-copy-capture-v1 (auto)","output":"hh-chromium-52w2dm","geometry":[1280,720],"mode":"raw","warmup":10,"frames":200,"frame_ms":{"min":118.647,"p50":187.814,"mean":205.548,"p95":298.366,"max":361.436},"fps_sustained":4.865,"wall_ms":41109.705,"cpu_ms_per_frame":1.652,"bytes_p50":null,"persistent":true,"rss_peak_mb":746.0,"rss_peak_kb_raw":763904,"system_cpu_ms_per_frame":408.6,"compositor_cpu_ms_per_frame":194.3,"machine_cpu_pct":198.8,"frames_bytes_allocated_p50":3686400,"notes":"grim-rs 0.2.0 via Grim::new() (auto-detect); copy target is a wl_shm buffer created inside each capture_* call, not a dma-buf the producer imports. Cursor not overlaid (the frozen argv has no cursor flag; hhstream-live's grim command passes -c). JPEG is Grim::to_jpeg, quality 80, the crate default and the live-stream setting. Raw timer stops when capture_* returns, which is after the crate copies the mapping into a Vec and swizzles XRGB/ARGB to RGBA (and applies any output transform or y-invert). The public API does not expose the mapped buffer before that, so raw includes work the contract says to leave outside the timer. Percentiles are linear interpolation on the sorted samples. cpu_ms_per_frame is getrusage(RUSAGE_SELF) utime+stime over the measured loop only, divided by frames. Frame buffers are freed between frames; that free is inside wall_ms and outside frame_ms. Outputs: Virtual-1@0,0 1280x720 scale 1; hh-chromium-gqkxz3@1280,0 1280x720 scale 1; hh-chromium-52w2dm@2560,0 1280x720 scale 1. No GPU on this machine; dma-buf was not used and is not claimed. This process does not move the pointer and does not start or stop a session; the pixels are whatever the compositor is already showing. Last warmup frame nonzero bytes: 3679565."}
Loaded 100 of 346 files, more files were not shown because too many files have changed in this diff. Show more